• Save
FCC Broadband Access Survey
Upcoming SlideShare
Loading in...5
×
 

Like this? Share it with your network

Share

FCC Broadband Access Survey

on

  • 2,531 views

The Federal Communications Commission's study on U.S. broadband access.

The Federal Communications Commission's study on U.S. broadband access.

Statistics

Views

Total Views
2,531
Views on SlideShare
2,524
Embed Views
7

Actions

Likes
1
Downloads
0
Comments
0

2 Embeds 7

http://community.pathoftheblueeye.com 6
http://www.community.pathoftheblueeye.com 1

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

FCC Broadband Access Survey Document Transcript

  • 1. OBI WOrkIng PaPer SerIeS nO. 1 BroadBand adoption and Use in america OBI WOrkIng PaPer SerIeS nO. 1 By John B. horrigan, Ph.D. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 1
  • 2. OBI WOrkIng PaPer SerIeS nO. 1 2 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 3. OBI WOrkIng PaPer SerIeS nO. 1 ➤ Among low-income Americans—those whose annual Summary Of fIndIngS household incomes fall below $20,000—broadband adoption stands at 40 percent. the Federal communications commission’s October- african-americans and Hispanics trail the average in november 2009 survey finds that nearly two-thirds (65 broadband access, although gaps have narrowed since percent) of american adults use high-speed Internet con- early 2009. nections to go online from home. ➤ 59 percent of African-Americans have broadband at The FCC conducted a survey of 5,005 Americans in October home. and November 2009 in an effort to understand the state of ➤ 49 percent of Hispanics (English and Spanish speaking) broadband adoption and use, as well as barriers facing those have broadband at home. who do not have broadband at home. The main findings are: ➤ For Hispanics who took the survey in Spanish, broad- band adoption is only 20 percent. ➤ 78 percent of adults are Internet users, whether that ➤ For Hispanics who opted to take the survey in English, means broadband, dial-up, access from home or access 65 percent have broadband. from someplace other than home. ➤ 74 percent of adults have access at home. These figures represent increases from levels registered in ➤ 67 percent of U.S. households contain a broadband user surveys conducted in early 2009 by the Pew Research Center, who accesses the service at home. which found in April that 46% of African Americans and 40% ➤ 65 percent of adults are broadband adopters. The dis- of Hispanics (English and Spanish speaking) used broadband crepancy of two percentage points between household at home. and individual home use is because some survey respon- dents are nonbroadband users but live with someone Some 42 percent of americans with disabilities have broad- who, at home, is. band at home. ➤ 6 percent of Americans use dial-up Internet connections The FCC survey asked adults a series of six questions to as their main form of home access. determine whether a respondent should be classified as having ➤ 6 percent are Internet users but do not use it from home; a disability. Nearly one-quarter (24 percent) responded “yes” they access the Internet from places such as work, the to at least one of the questions, indicating they have a disability; library or community centers. their broadband adoption rate is two-thirds the national aver- age. Looking at the data differently, 39 percent of all Americans For the purposes of this report, home broadband users are without broadband have some type of disability. those who said they used any one of the following technologies to access the internet from home: cable modem, a DSL-enabled Senior citizens (those over the age of 65) continue to trail phone line, fixed wireless, satellite, a mobile broadband wire- the national average in broadband adoption with a 35 per- less connection for your computer or cell phone, fiber optic, cent broadband-at-home penetration rate. T-1. In other words, home broadband users opt in to that classi- ➤ Nearly half (48 percent) of senior citizens are Internet fication through a survey question not by adhering to definition users, regardless of connection type. of broadband by speed that might be read to them. On average, americans pay nearly $41 per month for broad- the main dividing lines for access are along socioeconomic band service, but half of those who receive their broadband dimensions such as income and education. in a bundle with other services cannot identify the Internet ➤ 46 percent of adults whose highest level of education is portion of their bill. a high school degree are broadband users at home; 82 ➤ When asked the level of their monthly Internet bill, to the percent of adults who have attended or graduated from nearest dollar, broadband users, on average, cited a figure college are broadband users at home. of $40.68 ➤ 52 percent of Americans in households with annual ➤ Most users (70 percent) receive broadband bundled with incomes of $50,000 or below have broadband at home, other services. Of these, 50 percent could specify their compared with 87 percent of those in households with bundled monthly price: $37.70. incomes above that level. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 3
  • 4. OBI WOrkIng PaPer SerIeS nO. 1 ➤ Among broadband users who subscribe to a stand-alone three pitfalls cited are less active online than those who do not high-speed service at home (i.e., they do not have ser- share those views. These relationships do not show cause and vice in a bundle), the average monthly bill reported was effect, but they do suggest how attitudes influence the way $46.25. people engage with the Internet. Broadband users overwhelmingly view the social aspects of non-adopters are almost 50 percent more likely than the Internet as very important to them, while watching tV, broadband users to say they believe it is too easy for per- videos or movies online and playing games were deemed sonal information to be stolen online. less important. ➤ By a 57 percent to 39 percent margin, non-adopters of When asked what types of online applications are important broadband at home say they strongly agree that it is too to them, here is what broadband users said were very important easy to have their personal information stolen online. to them. ➤ 68 percent cited the Internet’s capacity to ease This is one factor linked to their lower likelihood of adop- communication with family and friends. tion. Although this concern is not necessarily a causal factor ➤ 39 percent said the ability to use the Internet to keep up behind non-adoption decisions, there is a significant positive with community news. correlation between high levels of worries about personal pri- ➤ 35 percent mentioned the ability to share content, such vacy and non-adoption. as photos, videos or text. ➤ 24 percent cited online shopping. Broadband users exhibit varying degrees of understanding ➤ 10 percent said watching television shows, videos of digital concepts. that, in turn, influences what they do or movies online. online. ➤ 9 percent identified playing games online is very Respondents received a series of questions asking how well important. they understood various terms relating to computers and the Internet; such questions serve as proxy measures for people’s although most users of broadband at home have positive online skill levels. What follows is the share of broadband users perspectives on the Internet, worries about inappropriate who say they understood very well the listed terms: online content and security of personal information can ➤ 61 percent—refresh or reload put a damper on online activity. ➤ 44 percent—operating system Most broadband users see the Internet as a tool for learning ➤ 42 percent—Internet browser cookie and productivity. ➤ 41 percent—JPEG file ➤ 81 percent strongly agree that the Internet is a valuable ➤ 40 percent—spyware or malware resource for information and learning. ➤ 16 percent—widget ➤ 74 percent strongly agree that it is important for children to learn to use the Internet. Those with greater understanding of these terms are heavier ➤ 56 percent strongly agree that people can be more online users. Some 29 percent of broadband users said they productive using the Internet. did not understand any of the listed terms very well, while 24 percent understood five or six of the terms very well. On aver- Nonetheless, broadband users view the Internet as having age, the former group engaged in about half as many online pitfalls. activities as their more informed counterparts. Again, a causal ➤ 56 percent strongly agree that too much pornography relationship is not suggested here, but the findings indicate and offensive material is online. how skills can shape how heavily they use the internet. ➤ 39 percent strongly agree that it is too easy for personal information to be stolen online. People take advantage of multiple devices and services to ➤ 24 percent strongly agree that the Internet is too go online, but these are usually supplementary access paths dangerous for children. for them. ➤ 30 percent of American adults have used a handheld Users with positive outlooks about the Internet tend to be device (e.g., cell phone or smart phone) to access the more active online (as measured by how many online activi- Internet.1 ties they do) than those with less upbeat views on the Internet. ➤ This behavior is more prevalent among minority groups. Similarly, broadband users with strong concerns about the 4 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 5. OBI WOrkIng PaPer SerIeS nO. 1 ➤ 39 percent of African-Americans have accessed the there are three primary reasons why the 35 percent of Internet with a handheld device. non-adopting americans do not have broadband: cost, lack ➤ 39 percent of Hispanics have accessed the Internet of digital literacy and broadband is not sufficiently relevant with a handheld device. for them to purchase it: ➤ 84 percent of those who use the Internet via a handheld ➤ 36 percent of non-adopters cite cost as the main rea- device have broadband at home. That figure is somewhat son they do not have high-speed Internet at home. This lower for African-Americans (78 percent) and Hispanics breaks out in the following ways: (68 percent). ➤ 15 percent say the price of the monthly bill is too much ➤ 15 percent of all Americans use a mobile wireless broad- for them. band service with their laptop computers.2 ➤ 10 percent say the cost of a computer is too much. ➤ The vast majority (92 percent) of wireless broadband ➤ 9 percent say they do not want a long-term service users have wireline broadband at home. contract or cannot afford the installation fee. ➤ Broadband users also access the Internet from places ➤ 2 percent cite a combination of these reasons. other than home. ➤ 22 percent of non-adopters cite factors pointing to lack ➤ 67 percent have at some time used the Internet at the of digital literacy as the main reason they are not online. homes of friends or family. These include people who are not comfortable with com- ➤ 62 percent have used the Internet at work. puters or, for non-internet users, are “worried about all ➤ 33 percent have used the Internet at a public library. the bad things that can happen if I use the Internet.” As ➤ 30 percent have used the Internet at school. people who cite digital literacy as barrier tend to be older ➤ 13 percent have used the Internet at a community (the median age is 62), concerns about the safety of the center. online environment is understandable. ➤ 5 percent have used the Internet at a place of worship. ➤ 19 percent of non-adopters do not have broadband because they question its relevance to their lives. They thirty-five percent of americans do not use broadband at do not believe digital content is sufficiently compelling home. they fall into three categories, each with distinct to justify getting it. Specifically, these non-adopters say demographic characteristics. the Internet is a “waste of time,” do not think there is ➤ 22 percent of adults are not Internet users. They are the anything worth seeing online and (for dial-up users) say oldest non-adopting group, with a median age of 60, and they are content with their current service. Dial-up users include the highest share of Hispanics (at 20 percent). make up a disproportionate share of those citing lack of Some 84 percent have high school degrees or less and half relevance as a barrier. live in households with annual incomes of $30,000 per year or less. non-adopters concerned with cost would be willing to pay, ➤ 6 percent of adults have dial-up connections at home. on average, $25 per month for broadband. Nearly one-third (29 percent) of them live in rural areas, Non-adopters who cited the monthly cost of broadband as a twice the rate for broadband users, and one-quarter reason they did not have service received a follow-up question (23 percent) have college degrees, which modestly lags asking them to estimate how much they would pay for service. the national average. The median age is 53. Among this group: ➤ 6 percent of adults are Internet users, but do not have ➤ 52 percent were able to provide an estimate; it averaged access from home. The group is relatively young (the $25 per month. median age is 38) and female (59 percent), but they ➤ 28 percent answered “don’t know” to this question. tend to have low incomes and low levels of educational ➤ 20 percent said they were not willing to pay anything for attainment.3 broadband. Of these non-adopters, 12 percent say they cannot get broad- Some 65 percent of those who estimated their willingness to band where they live. This translates into a 4 percent share pay (WTP) cited a figure of $20 per month or more. of Americans—on the basis of their reports on infrastructure ➤ If all of those who cited this figure had a service offering availability in their neighborhood—who say they are unable to at $20 before them and took service, broadband adoption obtain broadband because it is not available. This means that would be 6 percentage points higher in the United States. 31 percent of all Americans can get service but do not. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 5
  • 6. OBI WOrkIng PaPer SerIeS nO. 1 Nearly all (91 percent) who provided a WTP estimate identi- Together, users in these two groups with direct experience fied a figure of $10 per month or more. with broadband equate to 24 percent of the non-adopting ➤ If all of those who cited this figure had a service offering population, or 8 percent of all adults. at $10 before them and took service, broadband adoption would be 8 percentage points higher in the United States. “Proxy Internet” use is evident for 22 percent of non-In- non-adopters have a variety of outlooks on the Internet ternet or “not-at-home” users who live with someone with and information and communications technologies (Icts). online access. among these users: Some non-adopters, notwithstanding worries they may have ➤ 16 percent ask the Internet user in the home to carry out about the Internet and their capacity to use it, have a hopeful an online task at least once a week. outlook on the benefits it may confer. Many are users of ICTs, ➤ 20 percent ask the Internet user in the house to carry out just not broadband at home. Among nonbroadband adopters: an online task about once a month. ➤ 80 percent have premium television, i.e., either satellite ➤ About half of this “proxy access” is done using a broad- or cable. band connection in the home. ➤ 70 percent have cell phones. ➤ 42 percent have at least one working computer at home. When it comes to outlooks toward the Internet and levels of ownership of Ict products, non-adopting americans fall With respect to attitudes, most express high degrees of worry into four categories, each with different barriers to broad- about the privacy of their personal information and inappropri- band adoption. ate online content: digitally distant non-adopters make up 10 percent of the ➤ 65 percent strongly agree there is too much pornography general population, and members of this group do not see the and offensive material on the internet. point of being online. Few in this group see the Internet as a ➤ 57 percent strongly agree that it too easy for their per- tool for learning and most see it as a dangerous place for chil- sonal information to be stolen online. dren. This is an older group (the median age is 63), nearly half ➤ 46 percent strongly agree that the internet is too danger- are retired and half say that either lack of relevance or digital ous for children. literacy are barriers to adoption. The digital Hopefuls make up 8 percent of the population. At the same time, many non-adopters see the upsides to They like the idea of being online but lack the resources for ac- online access cess. Few have a computer and, among those who use one, few ➤ 59 percent strongly agree that the Internet is a valuable feel comfortable with the technology. They are most likely to tool for learning. cite cost as a barrier to adoption, with affordability of the com- ➤ 54 percent strongly agree that it is important for children puter playing an important role. They are also more likely than to learn to use the Internet. average to say digital literacy is a barrier. Demographically, ➤ 37 percent strongly agree that people can be more pro- this group is heavily Hispanic (26 percent), has a high share of ductive using the Internet. African-Americans (20 percent) and is low-income. The digitally uncomfortable make up 7 percent of the One-quarter (24 percent) of non-adopters have had expe- population, and are the mirror image of the Digital Hopefuls; rience with broadband, meaning they once had service at they have the resources for access but not a bright outlook home or have used it at work or someplace else: on what it means to be online. Nearly all of the Digitally ➤ 17 percent of all non-Internet users had home access in Uncomfortable have computers, but they lack the skills to use the past. them and have tepid attitudes toward the Internet. This group ➤ Of this group, 49 percent said they had home high- reports a variety of barriers, including cost, lack of available speed Internet access. infrastructure where they live, low perceptions of the Internet’s ➤ 46 percent of dial-up or “not-at-home” Internet users relevance and low digital literacy. have used a broadband connection from somewhere other The near converts, who make up 10 percent of the popu- than home, such as at work, school or a friend or family lation, have many of the same characteristics of broadband member’s house. adopters. They have high rates of computer ownership, positive 6 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 7. OBI WOrkIng PaPer SerIeS nO. 1 attitudes about the Internet and are, relative to other non- take a class for credit online (37 percent versus the 26 percent adopters, youthful (at a median age of 45). Many are dial-up average). With respect to barriers to adoption, 42 percent of or “not-at-home” users, and monthly access cost is the largest African-Americans say cost is the main reason they do not have reason for non-adoption among this group. broadband. looking across various population segments reveals dif- ferences in what they do online and what keeps them from Hispanics: Half (49 percent) of Hispanics have broadband at having broadband. home. While young African-Americans have broadband access on par with the average, “under 30” Hispanics trail the aver- Families: Parents with minor children at home are more age for their age cohort, 57 percent versus 75 percent. Those likely than average to have broadband at home, by a 75 percent Hispanics with broadband use it for downloading and stream- to 65 percent margin. The strong majority (87 percent) has at ing music (69 percent have done this, in comparison with the least one working computer at home, leaving a 13-percentage- 52 percent average). They also find broadband very important point gap between adopters and those with a functioning for keeping up with news about their community; 52 percent computer. Among those without broadband, nearly half (48 say this compared with the 39 percent average. As to adoption percent) cite cost as the main barrier to adoption, with 24 per- barriers, 52 percent of Hispanic non-adopters cite cost, equally cent specifically pointing to the level of the monthly access fee. split between those who point to the monthly fee and those who say they cannot afford a computer. Low-income Americans: Some 40 percent of low-income Americans (with annual household incomes at $20,000 or People with disabilities: Of the 24 percent of respondents below) have broadband, compared with 91 percent among those who have some sort of disability, 42 percent have broadband. living in homes with annual incomes above $75,000. Low-in- The online activities of broadband-using people with disabili- come broadband users are more likely than well-off broadband ties are narrower in scope than others; that is, they do fewer users to look for or apply for a job online – by a 77 percent to 60 things online. That may reflect difficulties some people with percent margin. Not surprisingly, 47 percent of this group iden- disabilities have in using the devices to get online or interacting tifies cost as the biggest barrier to broadband adoption. That is with Web pages. Barriers for non-adopting people with disabili- twice the rate of the (relatively few) upper-income nonbroad- ties do not differ significantly from the average. band adopters. Rural Americans: Fifty percent of rural residents have African-Americans: Although African-Americans trail broadband, a rate that reflects in part the older and less the average in overall broadband adoption, the adoption gap wealthy rural population but also the lack of available infra- does not exist for adults under age 30; some 75 percent of structure. One in 10 rural non-adopters say they cannot get African-Americans in this younger age group have broadband, broadband where they live. That is more than twice the aver- which equals the average. By contrast, 21 percent of African- age. Rural Americans with broadband, meanwhile, are as active American senior citizens have broadband; that compares with as their urban and suburban counterparts in using the Internet the 35 percent average. African-American broadband users are for shopping and taking classes online, suggesting that they use highly likely to have used their broadband connection to look broadband as a way to virtually access the benefits associated for or apply for a job; 83 percent versus the 60 percent aver- with urban or suburban living. age. African-Americans with broadband are also more likely to Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 7
  • 8. OBI WOrkIng PaPer SerIeS nO. 1 8 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 9. OBI WOrkIng PaPer SerIeS nO. 1 TaBle Of COnTenTS Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 I. Adoption: who uses broadband at home—and who doesn’t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 Who uses broadband at home–and who doesn’t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .13 Dividing lines on broadband access. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 Adoption patterns for key population segments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 How people connect to the Internet at home. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .14 Monthly cost of service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 II. Online behaviors: what shapes them and what matters to users . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16 Online activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16 Attitudes about the Internet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .16 Skills: Understanding information and communications technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17 The activities people say are most important . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18 Why new users get online . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19 III. Alternative access: different place, different platforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 The places people access the Internet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Use of the handheld for online access and other purposes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21 Mobile Wireless Broadband . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 What people pay for cell phone service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 IV. Non-adopters: who they are and the barriers they face . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Overview of non-adopters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Past and proxy broadband use for adopters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 The role of disabilities in non-adoption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Why non-adopters do not have broadband . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 What non-adopters are willing to pay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Summary of most important reasons for non-adoption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 When respondents can cite multiple reasons for not adopting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31 Segmenting the population of non-adopters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31 V. Focus on key population groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Families. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Low-income Americans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 African-Americans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Hispanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 People with disabilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Rural Americans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39 Appendix: Tables on demographics and use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 9
  • 10. OBI WOrkIng PaPer SerIeS nO. 1 10 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 11. OBI WOrkIng PaPer SerIeS nO. 1 For the survey, the margin of error based on results based InTrOduCTIOn on the entire sample of 5,005 is plus or minus 1.6 percentage points. For results on 2,671 home broadband adopters, the The Federal Communications Commission in the fall of margin of error is plus or minus 2.1 percentage points; for the 2009 fielded a national survey of Americans under author- oversample of non-adopters (n=2,334) the margin of error is ity granted by the Broadband Data Improvement Act (BDIA). plus or minus 2.2 percentage points. Finally, results based on The BDIA, which became law in October 2008 but needed all internet users (n=3,555), the margin of error is plus or mi- and received funding approval under the American Recovery nus 1.8 percentage points. Interviewers conducting the survey and Reinvestment Act (ARRA), requires the FCC to conduct provided a Spanish-language option for respondents wishing to periodic surveys that explore the broadband experience of take the survey in Spanish. American consumers. Specifically, the BDIA directs the survey The FCC’s survey results compare favorably to a similar to probe the types of technology people use for their home survey conducted in the fall of 2009 by the Census Bureau broadband service, what they pay, the applications they use for the National Telecommunications and Information and, for those who do not use broadband, the barriers to adopt- Administration, which found that 63.5 percent of American ing broadband. households used high-speed internet to go online.4 This paper reports the national findings from the random digit dialing telephone survey that interviewed 5,005 adult What is neW about this survey Americans, 2,671 of whom are broadband-at-home users. The The survey is unique for the FCC, in that it is the first time that remaining 2,334 either do not have broadband at home, report the Commission has conducted its own survey of end-users of that they do not use the Internet or say they are Internet us- the Internet. The Commission, using data gathered from FCC ers but without access at home. The number of non-adopters Form 477 (a semi-annual reporting document for telecom- surveyed represents an oversample of this population; this was munications entities) has previously reported on the number done to facilitate statistical analysis of non-adopters. of broadband lines in service to end users based on carrier-re- The Methodology section in the report’s appendix provides ported data at the Census tract level.5 This survey, as it involves detail on the sampling design for this survey, and also discusses interviewing adults, presents adoption data based on what such issues as response rate and use of cell phone numbers in users say about how they access the Internet at home. the sample. Cell phone numbers were used in the sample in The survey is distinct in a broader sense owing to its focus recognition of the fact many Americans do not have landline on non-adopters of broadband at home. There have been other telephone service at home, but rather rely only on their cell national surveys that have examined why people do not use phones. Overall, interviewers conducted 31% of the surveys the Internet, most notably the Pew Internet & American Life with respondents who were on a cell phone; 13% said they were Project’s 2002 survey on why people did not use the (mainly) “cell phone only” households, while 18% had a landline, yet dial-up Internet.6 Other Pew Internet research has analyzed were contacted on their cell phone. non-adoption of broadband based on a few questions in na- It is also worth noting that the data presented throughout tional surveys.7 is weighted to take into account characteristics of the sample, Finally, the FCC survey develops segments of non-adoption. including the oversample of non-adopters. The weighting This segmentation permits non-adopters to be placed in cat- corrects for the fact that, in the figures noted in above, 2,671 egories that yield insights into the reasons people do not have adopters does not represent 65% of 5,005 respondents, the broadband—and possible solutions that might be targeted at home broadband adoption figure highlighted at the outset. In non-adopters. The British regulator Ofcom undertook a similar results reported within, percentage figures represent weighted segmentation of non-adoption8 but this is the first time na- results; the number of cases reported for various sub-samples tional data collected by the U.S. government has been examined are actual number of cases. in that way. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 11
  • 12. OBI WOrkIng PaPer SerIeS nO. 1 12 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 13. OBI WOrkIng PaPer SerIeS nO. 1 This means that 22 percent of Americans are not Internet I. adOPTIOn: WhO uSeS users (1 percent of respondents did not know the type of Internet connection they have at home and thus are not classi- BrOadBand aT hOme— fied as broadband, dial-up or “not at home” online users). Although broadband access and reasons for not having it and WhO dOeSn’T are the principle themes of the survey, respondents were asked about several other assets pertaining to information and com- munications technology (ICT), such as use of computers, TVs Who uses broadband at home—and Who doesn’t and cell phones. Specifically: The October-November FCC survey finds that a strong major- ➤ 86 percent of Americans have a cell phone. ity of Americans use the Internet. Among adults, 78 percent ➤ 65 percent have cable TV at home. are Internet users, meaning they have access from home or ➤ 66 percent have a desktop computer at home. someplace other from home, and with any type of connection ➤ 52 percent have a laptop computer at home. (i.e., dial-up or broadband). The findings about access: ➤ 29 percent have satellite TV. ➤ 65 percent of adults use broadband at home. Combining cable and satellite TV, as well as desktop and lap- ➤ 67 percent of households have broadband. The survey top adoption, yields the following figures for home computer asked non-adopters whether another person in the house and home premium TV use: uses the Internet with a home high-speed connection. ➤ 86 percent of Americans have either cable TV or satellite Affirmative answers to that question add two additional TV. percentage points to overall penetration. ➤ 79 percent of Americans have either a desktop or laptop ➤ 6 percent of Americans have dial-up Internet connections computer at home. at home. ➤ 6 percent of Americans are Internet users but do not ac- The appendix to this paper provides detail on adoption pat- cess the Internet from home. terns for each of these technologies across demographic and socio-economic categories. Exhibit 1: 100 Broadband Adoption 93 90 86 by American Adults by 84 Socio-Economic and 80 77 75 74 Demographic Factors 70 68 69 American adults 64 59 59 60 55 50 49 50 40 40 35 30 24 20 10 0 High school graduate Some college College + Less than $20K $20–40K $40–75K $75K + Non Rural Rural 18–29 30–49 50–64 65+ White Black Hispanic* Less than high school *Hispanics includes both English and Spanish-speaking Hispanics Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 13
  • 14. OBI WOrkIng PaPer SerIeS nO. 1 dividing lines on broadband access take the survey in Spanish; among these Spanish speakers, Broadband adoption is not distributed evenly in the population broadband adoption was 20 percent. (See Exhibit 1). The differences in adoption break out promi- nently along two particular dimensions, which are not mutually People with disabilities: The FCC survey finds that 42 per- exclusive: cent of those who identify themselves with disabilities has broad- band at home. A July 2003 survey from the Pew Internet Project Education:Among adults, roughly half of Americans have showed that 10 percent of those with disabilities had broadband had some college experience (even if they have not or did not at a time when 18 percent of Americans had broadband.9 graduate). Among those who have some college experience, 82 The FCC survey sampled all Americans, but for several percent have broadband at home, compared with 46 percent of racial or ethnic categories, there were not enough respondents those whose highest level of educational attainment is a high to draw statistically reliable inferences. For Asian-Americans, school degree. American Indians, and Alaskan natives, the sample yielded fewer than 100 respondents in each group. The first two groups Income: Americans in the lower half of the income distri- in particular have a sizable population that may not speak bution demographic, meaning an annual household income of English or that have low telephone penetration rates. Because $50,000 per year or less, are much less likely to have broadband of that and the small sample of respondents, it is inadvisable to at home than those with higher incomes. Among respondents report results. reporting household incomes of $50,000 or less, 52 percent have broadband at home, far below the 87 percent adoption hoW people connect to the internet at home rate for those above that income threshold. The survey asked respondents what type of Internet connec- tion they use at home. The question, which is the same one adoption patterns for key population segments used in surveys conducted by the Census, allowed respondents Surveys conducted by the Pew Research Center’s Internet & to pick any of the categories listed that the respondent believed American Life Project found that broadband adoption within described his home broadband connection. In other words, two minority groups, African-Americans and Hispanics, lags respondents could pick more than one type of home broadband the average and, for African-Americans, has grown very little connection from the list (See Exhibit 2). from 2007 to the beginning of 2009. The FCC survey shows The answers underscore the difficulty in asking people some breaks from those patterns. about the technological dimensions of their home Internet access points. Although it is possible that some consumers african-americans: FCC figures found 59 percent of have more than one home high-speed connection, these figures African-Americans had broadband at home in the October- suggest that most Americans have two broadband connec- November survey, up considerably from 46 percent in Pew’s tions. It is likely that a significant number of consumers do not April survey. know the details of their home Internet connections. Some, for instance, may confuse a wireless home network working hispanics: Slightly less than half (49 percent) of Hispan- off wireline broadband with a “fixed wireless provider.” Given ics have broadband at home in the FCC survey compared with the opportunity to pick more than one category, in conjunc- 40 percent of Hispanics in a Pew Research Center April 2009 tion with uncertainty over how they connect to the internet survey. In the FCC survey, 3 percent of respondents chose to Exhibit 2: Cable modem 58 At home, what do you DSL-enabled phone line 44 now use to connect to the Mobile broadband wireless connection for your computer or cell phone 44 Internet ( figures as a Fixed wireless provider 29 percent of home Internet Dial-up telephone line 12 users) Satellite connection 10 A fiber optic connection such as FiOS 10 T-1 connection 5 Source: Federal Communications Commission survey of 5,005 adult Americans, October-November 2009. 14 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 15. OBI WOrkIng PaPer SerIeS nO. 1 at home, some respondents inaccurately chose more than one bundle, whether the respondents knew the Internet portion category of connection type. and, if so, what the monthly Internet bill was. Notwithstanding the possible confusion reflected in the This careful approach to bundling has the advantage of strip- survey responses, it seems likely that the vast majority of home ping out those who do not know the Internet portion of their broadband access is wireline. In fact, estimates place wireless bundle before they are asked to quantify their monthly Internet home broadband access at 2 percent of homes—that would bill. Those who do not know are not asked to make an estimate. include fixed wireless or satellite service.10 However, by appropriately excluding these respondents, the The survey also made extensive efforts to determine overall sample of respondents offering cost estimates is re- whether a respondent had dial-up service at home or not. First, duced. One-third of broadband users did not give an estimate of interviewers asked respondents first whether they connect at their monthly broadband payments. home “through a slow-speed connection such as dial-up or…a For Internet users, bundling their bill with other services is high-speed, broadband connection.” Among home users (72 common. A majority of home broadband users (70 percent) say percent of respondents), 8 percent said they had a slow-speed their Internet service is bundled with another service in their connection. monthly bill, with 29 percent saying they receive a stand-alone Second, a follow-up question for the 8 percent who said bill for broadband. they had a slow speed connection asked them to confirm that Among those who bundle Internet with another service, 50 they “now use a dial-up connection to the Internet at home, percent said they knew what the Internet portion of the bill was, not a higher-speed connection.” Most (93 percent) who first a figure that was the same for broadband and dial-up users alike. said they had a slow-speed connection confirmed that they The survey finds that: now have dial-up at home. For the purposes of this report, only ➤ Overall, users report paying $40.68 per month for their those who were able to confirm in the second question that broadband Internet connections.11 they have dial-up at home are defined as home dial-up users. ➤ Those who report they take service in a bundle (among This amounts to 6 percent of the entire adult population. those who can report what the Internet portion of the With the questioning approach isolating home dial-up bundle is), the average bill is $37.70. users, home broadband users are defined, in this report, as ➤ Those with stand-alone broadband service report a respondents who chose any of the options listed in Exhibit 2 monthly bill of $46.25. that are home high-speed connections, that is, cable modem, ➤ Those use dial-up at home to go online report a monthly a DSL-enabled phone line, fixed wireless, satellite, a mobile bill of $22.98. broadband wireless connection for your computer or cell phone, fiber optic, T-1. Thus, home broadband users opt in to Several other sources gather data on what people pay for that classification through a survey question not by adhering to broadband and offer points of comparison. They are: definition of broadband by speed that might be read to them. ➤ The Pew Internet Project found in its April 2009 survey that users report an average monthly broadband bill of monthly cost of service $39. As the BDIA directs, the survey also asked consumers what ➤ TNS Telecoms, through analysis of consumer bills, 90 they pay per month for home broadband service. This question percent of which are bundled offerings, finds an average presents another challenge for respondents, separate and apart broadband bill of $34.50. from knowing the details about home connection technology. ➤ Telogical Systems examines stand-alone nonpromotional Many consumers have Internet service bundled with other offers of providers, and its figures show an average broad- services (such as telephone or cable TV), making it difficult for band bill of $46 per month. the respondent to know what portion of the total bill reflects the cost of Internet service. Telogical’s figures for stand-alone prices track reason- To address bundling, the survey asked consumers if they paid ably well with what respondents in the FCC survey say about for various services (cell phone, landline phone, Internet, cable stand-alone bills. Those in the FCC survey able to identify the TV, satellite or wireless broadband) in conjunction with other Internet component of their bundle cited an average figure services. If consumers said they paid for the Internet in a sepa- about $3 more than the bill analysis of TNS Telecoms, which rate bill, they were asked to tell interviewers what their average focuses mainly on bundled offerings.12 monthly bill was to the nearest dollar. If respondents said their Internet service was bundled with another service, follow-up questions addressed what the monthly bills were for the entire Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 15
  • 16. OBI WOrkIng PaPer SerIeS nO. 1 half received the other eight. Among the first set of activities II. OnlIne BehavIOrS: (designated with an asterisk in the chart), broadband users on average participated in five of them. For the second set, broad- WhaT ShaPeS Them band users did 3.7 of them, on average. Those averages are affected by a host of factors. Young and WhaT maTTerS people are generally more attuned to digital life, so it is no surprise that adults under the age of 30 are typically more ac- TO uSerS tive. Respondents in that group say, on average, that they do 5.9 and 4.6 of the online activities across the two sets. The better This section describes the kinds of online activities people educated also do more things online: College graduates say engage in, as well as two factors that influence intensity of use. they have done 5.3 and 3.9 activities, on average, of the activi- This involves analyzing: a) how attitudes toward the Internet ties listed. A table in the appendix details how online activities influence how much they do online; and b) how the level of break out along demographic dimensions. understanding of computers and the Internet influences what As influential as demographics may be in explaining the they do online. The section concludes by looking at what scope of what people do online, there are two other relevant people find most important about their online activities. factors: attitudes about the Internet (its strengths and hazards) and skills (using measures of people’s understanding of com- online activities puters and the Internet as a proxy for their level of skill). The Internet, as has been well documented, is a means for communication, collaboration and content creation directed attitudes about the internet at users, shoppers, social networkers, information seekers and Respondents were read a series of six statements about the people searching for entertainment. The survey asked Internet Internet—three about the Internet’s upsides and three about its users if they had ever engaged in particular online activities, less attractive qualities (See Exhibit 4). with the topics chosen to reflect the variety of activities people In general, broadband users who showed strong levels of con- may pursue online. cern about potential hazards online reported doing a narrower To economize on the survey length, half the respondents scope of online activities than respondents without such wor- received eight of the listed activities in Exhibit 3, and the other ries. Similarly, those with positive perspectives were more active Exhibit 3: All Internet Dial-up Broadband Online Activities of users users users American Adults Buy a product online* 78 56 83 ( figures as a % of users Get local or community news* 75 55 80 in each group) Visit local, state or federal government Web site* 75 53 79 Use a social networking site* 52 41 55 Submit a review for a product or service* 52 36 55 Download or stream music* 47 22 52 Upload or share content* 45 26 48 Play games online* 46 38 48 Get international or national news 73 54 77 Bank online 63 43 69 Get information about or apply for a job 57 39 60 Get advice from a government agency about a health or safety issue 50 39 54 Download or stream video 38 18 42 Post to own blog or group blog 23 7 26 Take a class online 22 8 24 Play complicated role-playing games online 14 9 14 Source: Federal Communications Commission survey of 5,005 adult Americans, October-November 2009. Draft final results. For broadband users, n=1,378 for activities marked by * and n=1,278 for other activities. For dial-up user, n=212 for activities marked by * and n=247 for other activities. 16 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 17. OBI WOrkIng PaPer SerIeS nO. 1 online. It is important to note that worries about online dangers It is too easy for my personal information to be stolen do not always translate into behavioral changes. People may reg- online: Those who strongly agreed with this did an average of ister concerns about identity theft, but this may not prevent many 4.6 and 3.4 activities, 11 percent lower than those who do not users from sharing personal information at online social network- share this view (across both sets of activities). ing or shopping sites. Overall, broadband users who strongly registered any of The following analysis shows how scope of activity varies by the three worries listed above (theft of personal information, respondents’ attitudes. Recall that half the sample was asked inappropriate content, online dangers for children) engaged about eight different online activities, and the other half asked in 4.8 and 3.5 of the online activities probed. The share of about a separate set of eight activities. To show the interaction of all broadband users who cited any of those three things as attitude and activity, the analysis below reports the average num- strong concerns is 70 percent and, when compared with the 30 ber of activities (across both sets) among those who strongly agree percent of broadband users who share none of these concerns with a specific attitudinal proposition. That is then contrasted, in strongly, they do 14 percent fewer online activities. These com- percentage terms, with those who did not share that view. parisons do not mean that the attitudes in question cause or are sole drivers of the differences identified. However, they do The Internet is too dangerous for children: Those who indicate that how people perceive the Internet shapes how they strongly agreed with this did an average of 4.3 and 3.1 activities, use it. It is worth noting that the differences identified above 19 percent lower than those who do not share this view (aver- are statistically significant when controlling for other respon- aged across both sets of activities). dent characteristics, such as age, race, education and income.13 It is important for children to know how to use the skills: understanding information and Internet: : Those who strongly agreed with this did an average communications technology of 5.2 and 3.8 activities, 16 percent more than those who do not The scope of activities people do online may also have to do share this view (averaged across both sets of activities). with the skills they bring to the online experience. Measuring the level of skill or literacy for a user is challenging in a survey The Internet is a valuable source for information and environment. Merely asking people to rate how well they can learning: As was the case for the importance of the Internet carry out a given task can be problematic, and research has found for learning among children, those who strongly agreed with that such self-assessment about skill is not the best predictor of this engaged in an average of 5.2 and 3.8 activities, 25 percent actual skill levels. A better predictor of actual skill comes from more than those who do not share this view (averaged across asking respondents to rate on a scale their understanding of vari- both sets of activities). ous concepts relating to the Internet or computers.14 The FCC survey borrowed this approach, asking Internet us- There is too much pornography and offensive materi- ers how well they understand six terms relating to the Internet al on the Internet: Those who strongly agreed did an average or computers (See Exhibit 5). of 4.7 and 3.4 activities, 11 percent lower than those who do not Not surprisingly, education is strongly associated with the share this view (averaged across both sets of activities). likelihood that people say they understand a particular concept very well (See Exhibit 6). People can be more productive using the Internet: An important consequence of low levels of digital literacy Those who strongly agreed with this did an average of 5.2 and is less engagement with online life. Those who say they under- 4.0 activities, 18 percent more than those who do not share this stand the queried concepts “very well” are likely to do more of view (across both sets of activities). the online activities discussed. This correlation is significant Exhibit 4: The Internet is a valuable source for information and learning 81 Public Attitudes It is important for children to learn how to use the Internet 74 about the Internet There is too much pornography and offensive material on the Internet 56 (% of broadband users People can be more productive using the Internet 56 who strongly agree) It is too easy for my personal information to be stolen online 39 The Internet is too dangerous for children 24 Source: Federal Communications Commission survey, October-November 2009, n=2,671 for broadband users. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 17
  • 18. OBI WOrkIng PaPer SerIeS nO. 1 when holding constant income, education, age and other fac- the terms very well doing, on average, more than twice as many tors that might also affect people’s probability of doing a range activities as those who understood none of them. of things online.15 Exhibits 7, 8 and 9 demonstrate this point by showing, for the activities people say are most important broadband users, how the number of online activities in which What people do online and how they value a particular activity they engage increases in tandem with the number of computer can differ for a number of reasons, often due to the context in and Internet terms they say they understand “very well.” For which an activity is undertaken. More Internet users have, at the first set of activities, broadband users who did not say they some point, bought something online (78 percent) than have understood any of the computer or Internet terms listed (29 uploaded content they have created (45 percent) or partici- percent of respondents) engaged in an average of 3.8 of the pated in a blog (23 percent). However, seeing a video of a child’s online activities about which they were asked. Those who un- birthday party may be more important to some people than derstood very well all of them (24 percent of respondents) did, buying a product online – especially if offline shopping oppor- on average, 6.1 activities. The differences are more pronounced tunities are abundant. Conversely, coming upon a hard-to-find for the second set of activities, with those who understand all of Exhibit 5: Very well Somewhat well Not too well Not well at all Broadband Users’ Internet browser cookie 42 30 12 15 Understanding of Spyware and malware 40 34 11 14 Computer and Internet Operating system 44 29 11 16 Concepts ( figures as a Refresh or reload 61 22 5 11 % of users) Widget 16 13 15 54 JPEG file 41 22 8 27 Source: Federal Communications Commission survey, October-November 2009; n=2,671 for home broadband users. Exhibit 6: Less than High school grad Some college College+ By Education, High School Broadband Users’ Internet browser cookie 23 32 44 50 Understanding of Spyware and malware 26 33 41 47 Computer and Internet Operating system 21 31 44 55 Concepts (% “very well” Refresh or reload 46 52 64 68 in each group) Widget 4 11 15 22 JPEG file 17 28 41 55 Source: Federal Communications Commission survey, October-November 2009; n=2,671 for home broadband users. Exhibit 7: Average number of online activities Average number of online activities Number of Online (first set of eight activities) (second set of eight activities) Activities and Digital None 3.8 2.5 Literacy One term 4.9 3.3 Two terms 5.1 3.5 Three terms 5.4 4.2 Four terms 5.6 4.0 Five or more terms 6.1 4.8 Source: Federal Communications Commission survey, October-November 2009; n=1,378 for home broadband users for first set of activities and n=1,293 for home broadband users for second set of activities. 18 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 19. OBI WOrkIng PaPer SerIeS nO. 1 good in the far- flung corner of the Web may be extremely Why neW users get online important if the buyer is a collector or shopping for a special What existing users value about online activities and what trig- occasion. gers new adoption can be two different things. To address this, Even though context is key to interpreting how people value the survey asked about new users’ motives for getting online. different online activities, the survey nonetheless sought to The level of Internet penetration has not grown substan- explore this by asking Internet users to assess the importance tially in recent years. As noted, the FCC survey finds that 78 of six different categories of online activities. Exhibits 8 and 9 percent of Americans use the Internet, whether that is at home show results for all users and broadband users when they were or somewhere else, dial-up or broadband. Surveys conducted asked to assess how important each online activity was to them. by the Pew Internet Project show that Internet penetration For most Internet users, the communicative aspects of the first exceeded 70 percent in early 2006, with that figure reach- Internet are most important to them, followed by keeping up ing 75 percent for the first time at the end of 2007. The FCC with news in the community and sharing content with oth- asked how long respondents have been using the Internet; just ers. Applications oriented to entertainment rate notably low, 6 percent report that they have been online for two or fewer perhaps because (for watching TV shows or movies) excellent years. Those who have been online for two or fewer years offline alternatives exist. Young broadband users are more received follow-up questions asking them to cite the reasons likely to cite entertainment applications or online entertain- they chose to get Internet access and then the most important ment as “very important” than average, with 20 percent and 13 reason for getting access. percent saying this. They are also more likely to say communi- As Exhibit 9 shows, most new users cite social reasons as a cating with family and friends (79 percent) and sharing content motive for beginning to use the Internet, with e-mail communi- (45 percent) are very important online activities. cation and content sharing as chief reasons. Accessing video or Exhibit 8: % as a share of broadband Internet users at Somewhat Not too Not important Very important Broadband Survey home important important at all Users on the Importance Making it easy to communicate with friends and 68 23 5 4 of These Activities family, even if they are far away Keeping up with the news in my community 39 39 11 10 Sharing content with others, such as photos, 34 38 15 13 videos or text Shopping online 23 41 19 18 Watching TV shows, movies and other video 10 26 24 39 online Playing games online 9 18 23 50 Source: Federal Communications Commission survey, October-November 2009; n=2,671 for home broadband users. Exhibit 9: % who cite as % who cite Broadband Survey reason most important reason Users on Their Reasons for Getting Online To e-mail and stay in touch with family and friends 61 31 (% among Internet To gain access to music, movies and other entertainment 46 9 users who have been To share my photos or videos with family and friends 41 3 online for 2 years or less) My children wanted Internet access 41 7 My children needed it for school 35 19 I needed it for school 27 10 An Internet provider made a special offer too good to pass up 21 2 My job required online access 20 6 Source: Federal Communications Commission survey of adult Americans, October-November 2009. N=92 for new users. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 19
  • 20. OBI WOrkIng PaPer SerIeS nO. 1 other Internet entertainment online also was a factor for many laptop computer) and Internet access from places outside the new users. A great offer by an Internet service provider was not home (often public access points). widely cited by new users. When pressed about the most impor- tant reason, social and educational uses are most prominent. the places people access the internet E-mail is the main reason for one-third of new users, while Although most online users have access from home, there are school—either for themselves or children—is the main reason plenty of places outside the household where people can go for another one-third of new users. It is worth pointing out the online. To examine how people access the Internet away from small number of cases (92) for the subset of new users who home, respondents received questions that tried to fix whether received these questions. they have gone online from someplace other than home and, if so, where. Exhibit 10 shows the places from which people ac- cess the Internet. III. alTernaTIve These results show that 93 percent of Internet users have some form of access at home. Yet, it is apparent people access aCCeSS: dIfferenT the Internet in other contexts as well—particularly at work or a friend or family member’s house. PlaCeS, dIfferenT Exhibit 11 indicates that a majority of Americans (59 per- cent) who access the Internet do so in more than three places; PlaTfOrm 17 percent of Americans go online in 5 or more places. Focusing on alternative access places that exclude home or Access to the Internet is portable and shareable. People may work, 57 percent of all Americans (or 73 percent of Internet take access with them or, if they cannot, they may access the users) have at one point gone online at some place other than Internet from different places. The survey looked at this in home or work, often at a friend or family member’s house. three contexts: handheld access, mobile broadband (from a Eliminating the latter from consideration, 39 percent of Americans (or 48 percent of Internet users) have gone online Exhibit 10 : All Internet Broadband at Not-at-home Dial-up at home Broadband Survey users home users Users on Where They Home 92% 100% 99% 7% Access the Internet Friend or family member’s house 65 67 48 67 Work 58 62 36 48 Public library 35 33 32 57 School 29 30 17 21 Community Center 14 13 9 25 Church 5 5 4 5 Number of cases 3,555 2,671 459 392 Source: Federal Communications Commission survey, October-November 2009. Exhibit 11: (% of Internet users) Broadband Survey One access point 17 Users on the Number of Two access points 24 Places They go Online Three access points 25 Four access points 17 Five or more access points 17% Source: Federal Communications Commission survey, October-November 2009. 20 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 21. OBI WOrkIng PaPer SerIeS nO. 1 at a library, school, community center or church. It is worth the share of the adult population with smart devices at 17 noting that these figures do not include those who may access percent in November 2009.16 The survey did not ask directly the internet at a local coffee shop. whether a respondent had a smart phone, but rather whether Exhibit 12 shows how incidence of using the internet at cell users did specific activities—some of which are likely to different access sites varies across demographic and socio- require having a smart phone. economic characteristics of the respondent. Exhibit 13 shows the findings from these questions, with a breakout by age, as young adults are more frequent users of use of the handheld for online access and mobile devices than their elders. other purposes Although basic communication functions dominate the With Internet access increasingly in people’s pockets, it is portfolio of cell phone activities, there is a substantial inci- important to probe the activities people pursue using their dence of information seeking, such as accessing Web pages or handheld devices. Fully 86 percent of Americans have a cell searching for directions. Twenty percent of cell phone users say phone. Estimates of the share of people with “smart phones” they have downloaded an application to their device, suggest- that permit online access vary, but Forrester Research places ing that at least that share has a smart phone; the actual level of Exhibit 12: Community (% of those surveyed) Public library School Church Use of Public Access center Points and Respondent Male 33 26 15 5 Characteristics Female 38 31 13 5 Parents with minor children at home 35 28 13 5 Those who report they have a disability 35 22 16 4 18–29 56 58 26 5 30–49 33 26 13 4 50–64 25 14 8 4 65+ 20 6 5 5 White (not Hispanic) 31 25 11 4 Black (not Hispanic) 51 36 21 6 Hispanic (English or Spanish speaking) 46 43 26 5 Less than high school 41 24 19 2 High school graduate 33 21 14 4 Some college 41 34 15 5 College + 31 31 11 6 Under $20K 55 41 25 4 $20K–$30K 44 30 20 6 $30K–$40K 41 33 13 5 $40K–$50K 30 23 12 4 $50K–$75K 28 24 10 5 $75K–$100K 31 27 11 5 Over $100K 28 30 10 4 Urban 38 33 16 5 Suburban 33 27 12 4 Rural 34 22 13 4 Source: Federal Communications Commission survey, October-November 2009. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 21
  • 22. OBI WOrkIng PaPer SerIeS nO. 1 smart phone usage is probably higher since not all smart phone In defining the scope of Internet use on a mobile handheld, users are likely to download applications. the activities chosen for inclusion in the definition are: sending As documented by the Pew Internet Project, use of mobile or receiving e-mail, accessing Web pages on the Internet and devices for non-voice data applications is particularly popular downloading an application to a cell phone. When those three among African-Americans and Hispanics.17 Unlike the Pew items constitute accessing the Internet on a cell phone, 36 survey, the FCC survey had a Spanish-language option, which percent of all cell users have used the Internet on their devices. means data reported here on mobile use among Hispanics Since not everyone has a cell phone, this means 30 percent of represents a more complete sample of the Hispanic population. all adults have gone online with a cell phone or smart phone. Still, African-Americans and Hispanics, consistent with prior As with the entire suite of handheld activities, minorities are research, are the most frequent users of the handheld device for more active in handheld Internet use than whites. This is how a variety of applications (See Exhibit 14). results break down across racial categories: Not all of the activities listed in Exhibit 14 constitute ➤ 39 percent of Hispanics and of African-Americans have Internet use on a handheld. Texting is not an application that used the Internet on their mobile handheld device. runs on the Internet while downloading maps and directions ➤ 27 percent of whites have accessed the Internet on a mo- may require Global Positioning System (GPS) capability. Not bile handheld. all downloading or streaming of music or video to a handheld A question to consider is if the differences are racial in requires Internet access. origin or whether they can be explained by other factors, such as age or income, which may vary systematically across racial Exhibit 13: All adults Age 18–29 Age 30–49 Age 50–64 Age 65+ Cell Phone Activity % with a cell phone 86 94 93 86 66 % as a share of those with cell phone Send or receive text messages 66 94 79 51 15 Send or receive pictures 52 77 61 38 14 Send or receive e-mail 26 40 30 17 5 Send or receive Instant Messages 28 42 35 20 8 Access Web pages on the Internet 28 48 32 15 5 Get a map or directions to another location 27 44 30 16 8 Download an application to your cell phone 20 36 24 9 3 Download or stream music or video 17 32 18 8 3 Source: Federal Communications Commission survey of 5,005 adult Americans, October-November 2009. Exhibit 14: Whites African-Americans Hispanics By Race, Cell Phone % with a cell phone 87 86 85 Activities % as a share of those with cell phone Send or receive text messages 63 75 75 Send or receive pictures 50 61 58 Send or receive e-mail 24 34 28 Access Web pages on the Internet 24 36 36 Get a map or directions to another location 23 36 34 Send or receive Instant Messages 23 47 47 Download an application to your cell phone 17 26 24 Download or stream music or video 13 28 25 Source: Federal Communications Commission survey of 5,005 adult Americans, October-November 2009. 22 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 23. OBI WOrkIng PaPer SerIeS nO. 1 categories. White adults in the sample are, on the whole, older access replacing home Internet connection. Rather—since than the African-American respondents and much older than most mobile Internet users also have broadband—it is an ad- the Hispanics. The median age for white adults in the sample ditional pathway to access. This pattern is not quite as strong is 49; for African-Americans it is 46, and for Hispanics it is 33. for African-Americans and Hispanics, but the general trend is, As seen above, younger respondents are heavier users of mobile nonetheless, the same for those groups. devices than older ones. While age, education and income are It is important not to confuse mobile Internet use with all indicators of the likelihood of using the Internet on a hand- mobile broadband. To be sure, some of those using their mobile held device, the racial or ethnic classification of the respondent device for online access do so via relatively fast 3G and WiFi also emerges as a predictor—even when other factors are held networks. However, it is difficult, using a survey, to determine constant. In other words, the differences that appear across what portion of mobile Internet users do this or how often. races are not simply due to age or educational level. Another issue is whether mobile Internet access is a sub- mobile Wireless broadband stitute for home access. African-Americans and Hispanics Another access pathway for people is mobile wireless broad- report lower levels of home broadband adoption than whites band service on a laptop. The survey asked laptop users the yet are more likely to use the Internet on their handheld device following question to probe mobile broadband use: (See Exhibit 15). “Do you use a service with your laptop computer that is Are lower home adoption rates attributable to mobile called wireless broadband, allowing you to access the Internet Internet use? The survey results indicate the answer is, for virtually anywhere? This is usually a service that you have to the most part, “no.” There is a positive correlation between pay a monthly fee for, either by itself or as part of another com- home Internet adoption and mobile Internet use—91 percent munications bill. This is NOT what is called Wi-Fi.” of mobile Internet users have Internet at home and 84 percent This access pathway appears at only modest levels in the have broadband. Yet while African-Americans and Hispanics general population. Among laptop users (52 percent of all are more likely to be mobile Internet users, most users in these adults), 28 percent said they used wireless broadband. As a per- groups have broadband at home. More than three-quarters centage of the entire adult population, this comes to 15 percent (78 percent) of African-American mobile Internet users have who use wireless broadband. broadband at home, and more than two-thirds (68 percent) of African-Americans and Hispanics who have laptops were Hispanic mobile Internet users have high-speed access at home. more likely than average to say they use such a service. Looking Viewed from the perspective of nonbroadband adopters, few at the racial breakdown: non-adopters access the Internet on their handheld devices, ➤ Among African-Americans with laptops, 36 percent re- though African-Americans and Hispanics do this at a higher port using a mobile broadband service. rate than average. Specifically, among those who do not go ➤ Among Hispanics with laptops, 30 percent report using a online at home with a broadband connection: mobile broadband service. ➤ 14 percent of all non-adopters have accessed the Internet ➤ Among whites with laptops, 26 percent report using a on their mobile devices. mobile broadband service. ➤ 20 percent of African-American non-adopters have ac- cessed the Internet on a mobile device. The context for these findings is important. First, these ➤ 25 percent of Hispanic non-adopters have accessed the figures apply to respondents with laptop computers—devices Internet on their mobile devices. that Hispanics and African-Americans are less likely to have. Overall, there is not a high incidence of mobile Internet (Some 54 percent of whites have a laptop, while 44 percent of Exhibit 15: (% of mobile Internet users) By Race, Mobile and Internet at home Broadband at home Home Internet Use White 94 89 African-American 92 78 Hispanic 78 68 Total 91 84 Number of cases 3,455 2,671 Source: Federal Communications Commission survey of 5,005 adult Americans, October-November 2009. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 23
  • 24. OBI WOrkIng PaPer SerIeS nO. 1 African-Americans and 43 percent of Hispanics do.) As a share ➤ 6 percent use dial-up Internet connections to go online of the entire population of each of these groups, mobile broad- from home. band use unfolds as follows: ➤ Among all African-Americans, 16 percent say they use a Overall, 35 percent of Americans do not use broadband at mobile broadband service. home. (One percent of respondents were not broadband us- ➤ Among all whites, 14 percent say they use a mobile broad- ers, but could not be placed in any of three categories above). band service. A small share—just 4 percent—say they cannot get service ➤ Among all Hispanics, 13 percent say they use a mobile where they live. This is an imperfect measure of infrastructure broadband service. availability, as it relies on the user accurately knowing about infrastructure in the neighborhood. This means that about one- Second, as noted, mobile broadband users are overwhelming- third (31%) of Americans do not have broadband at home but ly home broadband users; 94 percent of all mobile broadband could subscribe. However, the full cohort of non-adopters (35 users have broadband at home, figures that are only modestly percent) is included in our analysis and discussion. lower for African-Americans (91 percent) and Hispanics (84 Key demographic differences when contrasting all non- percent). Although some may be using their mobile broadband adopters with adopters are: connection as their principle home access means, those saying ➤ gender: 57 percent of non-adopters are women versus 49 they use mobile broadband were as likely as the average to say percent of home broadband adopters. they use DSL, cable modem service and other wireline means ➤ People with disabilities: 39 percent of non-adopters such as fiber. As with mobile Internet use, mobile broadband is have a disability, compared with 15 percent of adopters. mainly a supplementary broadband access pathway. ➤ college graduates: Just 11 percent of non-adopters have college degrees versus 37 percent of broadband users. What people pay for cell phone service ➤ age: 32 percent of non-adopters are age 65 or older ver- The survey asked cell phone users what their monthly bill was, sus 9 percent of adopters. and the question included a prompt for respondents to include ➤ Nearly two-thirds (65%) of non-adopters who are charges for texting or long-distance service. Only respon- senior citizens are women. dents who receive a stand-alone cell phone bill—80 percent of ➤ Income: 43 percent of non-adopters live in households respondents—received this question. The average figure they with annual incomes of $20,000 or less, compared with 17 gave was $92 per month. percent of home broadband users. By comparison, bill analysis from TNS Telecoms shows an ➤ rural: 24 percent of non-adopters live in rural areas average monthly cell phone bill of $99 per month. Many cell versus 13 percent of broadband adopters. users have more than one line on their bill; for TNS, the aver- age bill covered 1.9 lines. The FCC survey did not ask how many Exhibit 16 presents detail on the demographic and so- lines users had on their plan. cio-economic profiles for each of the three categories of non-adopters. Non-Internet users, the largest group, are older, lower-income and the least educated of the three. Online users Iv. nOn-adOPTerS: who do not have home access tend to be younger women, usu- ally with no more than a high school degree and low incomes. WhO They are and Dial-up users, though not as well off economically as broad- band users, are higher on the socioeconomic scale than the The BarrIerS They other groups of non-adopters. faCe past and proxy broadband use for non- adopters Although those classified as non-adopters are not broadband overvieW of non-adopters users at home today, this does not mean that all of them are com- Roughly one-third of Americans do not use broadband Internet pletely cut off from modern ICT goods or services. Many have at home. They fall into three categories of non-adoption: a working computer at home (discussed below) and 49 percent ➤ 22 percent of Americans do not use the Internet at all. uses computers occasionally, either at home, work or school. ➤ 6 percent use the Internet but do not have access Additionally, some have had experience with broadband. at home. Among dial-up users or “not-at-home” users, 46 percent have 24 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 25. OBI WOrkIng PaPer SerIeS nO. 1 used broadband at some point—maybe at work, perhaps a Internet users at home. Of these former users, 49 percent say friend’s house or possibly somewhere else. Among the wider broadband was the type of home connection they had. set of all non-Internet users, 17 percent in this group were once Combining these two sets of nonusers’ past broadband Exhibit 16: Broadband at Dial-up Internet users, Non-Internet The Demographics home at home not at home users of Non-adopters Gender Male 51% 49% 41% 45% Female 49 51 59 55 Families Parents with minor children at home 36 24 27 19 Disabled Those who report they have a disability 15 19 28 47 Age 18-29 25% 16% 39% 11% 30-49 39 27 34 20 50-64 26 33 20 28 65+ 9 24 8 41 Median Age 43 53 38 60 Race/ethnicity White (not Hispanic) 74% 75% 64% 63% Black (not Hispanic) 10 10 15 13 Hispanic (English or Spanish speaking) 10 10 15 20 Educational attainment Less than high school 5% 12% 8% 39% High school graduate 29 37 50 45 Some college 29 28 23 11 College + 37 23 19 5 Household income Under $20K 10% 16% 30% 32% $20K-$30K 7 13 16 17 $30K-$40K 9 11 18 7 $40K-$50K 9 15 7 8 $50K-$75K 16 14 11 6 $75K-$100K 15 8 6 1 Over $100K 20 5 3 2 Don’t know/refused 15 19 10 27 Community type Urban 31 24 22 31 Suburban 51 43 40 45 Rural 14 29 21 22 Number of cases 2,671 459 392 1,450 Source: Federal Communications Commission survey, October-November 2009. Total number of cases in this table is 4,972. This does not reflect the entire sample size of 5,005 because some responses did not permit classification into any of the categories. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 25
  • 26. OBI WOrkIng PaPer SerIeS nO. 1 experience yields reveals that one-quarter of non-adopters the role of disabilities in non-adoption have had some experience with broadband—meaning they once The largest group of nonbroadband adopters—non-Internet had broadband access at home or have used it at work, a friend’s users—has a high likelihood of having some sort of disability. house or some other place. Put differently, 8 percent of the Overall, 39 percent of non-adopters have some sort of disabil- general population have had some exposure to broadband but ity. The survey posed six questions, which, if any one yielded a do not presently have it at home. “yes” answer, classified the respondent as a disabled American As to prospects for moving nonusers to broadband, 20 per- (See Exhibit 17). cent of nonusers say they would be interested in getting online Collectively, 24 percent of respondents had an affirmative access. Among current “not-at-home” Internet users, 46 response to one of those questions and, for purposes of this percentwould be interested in getting home access, while 41 report, are classified as people with disabilities.18 Within this percent of dial-up users say they would be interested in getting group, 42 percent have broadband at home—much lower than broadband. Only 17 percent of dial-up users say they are very the 65 percent average. Some 56 percent of those with disabili- likely to get broadband in the next year. Similarly, among all ties are Internet users, below the 78 percent average. As the interested nonusers, broadband access is not a sure bet in the exhibit presenting the demographic and socioeconomic profiles near future. Out of all non-adopters, 26 percent say it is very of non-adopters shows, nearly half (47 percent) of non-Internet likely they will get broadband in the next 12 months, with an- users report having some sort of disability. other 28 percent saying this is somewhat likely. It is important to note that people with disabilities and There is also the phenomenon of proxy access, that is, non-adopters overlap quite a bit and in fact share many charac- nonusers who cohabit with another Internet user. Among teristics with non-adopters: They are generally older and have respondents who are non-Internet users or who do not use the lower incomes. For that reason, it is not a surprise that non- Internet at home, 22 percent live with someone who uses the adopters include a disproportionately high share of people with Internet at home. These nonusers often ask their online house- disabilities. Some of the difference in adoption rates is due to in- mates to carry out tasks online for them. One in six (16 percent) dividuals’ disabilities and some is due lower incomes, advanced say they ask their online housemates to do something online age or other factors associated with low adoption. for them once a week. Another 20 percent say they do this once a month, while an additional 24 percent say they do a few times Why non-adopters do not have broadband a year. The survey sought to determine with as much specificity as A portion of the proxy access that nonusers’ housemates possible why people without broadband choose not to have the carry out is broadband. Nonusers who live with someone who service at home. Respondents were read a list of possible rea- goes online from home were asked about the household’s sons for not having broadband and permitted to list as many as access. Some 21 percent say it is a dial-up connection, 47 per- they chose as barriers. They then received a follow-up question cent state it is some other way and 32 percent do not know. asking the most important reason they did not have broadband. Assuming that the response “some other way” reflects a home The question wording and choices offered varied slightly high-speed connection, these questions suggest that 2 percent across different non-adopter groups. Dial-up users, for ex- of the general population fall into the category of nonusers who ample, were offered the choice of “happy with current service,” live in a broadband-connected household. which is not a relevant choice for those who do not use the Internet at all. A number of the exhibits show how people responded to the question that allowed them to list multiple Exhibit 17: Do you have serious difficulty walking or climbing stairs? 12 Questions Used to Because of a physical, mental or emotional condition, do you have serious difficulty concentrating, 8 Determine Disability remembering or making decisions? (by Percentage) Are you deaf or do you have serious difficulty hearing? 8 Because of a physical, mental or emotional condition, do you have difficulty doing errands alone such as 6 visiting a doctor’s office or shopping? Are you blind or do you have serious difficulty seeing even when wearing glasses? 5 Do you have difficulty dressing or bathing? 2 Source: Federal Communications Commission survey of 5,005 adult Americans, October-November 2009. 26 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 27. OBI WOrkIng PaPer SerIeS nO. 1 reasons for non-adoption, then how they responded to the users. Some 12 percent are worried about risks online, and 14 question asking about the most important reason. percent cite factors that suggest they are unclear about the Internet’s relevance to them (with 7 percent saying the Internet Non-Internet users—22 percent of the adult population is a waste of time and 7 percent saying there is nothing they Those who do not use the Internet offer a range of reasons for want to see online). (See Exhibit 19) not being online. Nearly half cite the monthly cost, and similar Those who cited monthly cost as a barrier (47% in this cat- numbers point to comfort with computers or worries about egory of non-users) received follow-up questions on what they online content—that is, digital literacy looms as a barrier for would be willing to pay for service. The responses sorted into them. The affordability of the computer is also a prominent three categories: reason for nonusers (See Exhibit 18). ➤ 40 percent of nonusers provided an estimate and, among The survey permitted respondents to choose more than one this group, the average monthly figure they said they reason from the list of barriers to adoption. Fifty percent of would pay for Internet access at home was $26; the me- nonusers picked at least three reasons, and one-quarter (24 dian figure was $20. percent) cited at least five of the reasons listed. ➤ 31 percent said either they don’t know or refused to When asked the most important reason for not using the answer. Internet, no single reason stands out for nonusers. About 16 ➤ 29 percent said they were not willing to pay anything for percent cite lack of comfort in using a computer as the main access. barrier, but cost (either monthly fee, cost of computer or ac- tivation fee) is the main reason cited by 25 percent of dial-up Exhibit 18: Monthly cost is too expensive 47 Reasons Nonusers I am not comfortable using a computer 46 Do Not Use the Internet I am worried about all the bad things that can happen if I use the Internet 45 (by Percentage) The activation and installation fee to get service is too much 42 I cannot afford a computer 40 There is nothing on the Internet I want to see or use 35 The Internet is just a waste of time 33 I can access the Internet all I need to at work 14 It’s not available where I live 13 Source: Federal Communications Commission survey, October-November 2009; n=1,450 for nonusers. Exhibit 19: I am not comfortable using a computer 16 Main Reasons Nonusers I cannot afford a computer 14 Do Not Use the Internet I am worried about all the bad things that can happen if I use the Internet 12 (by Percentage) Monthly cost is too expensive 11 The Internet is just a waste of time 7 There is nothing on the Internet I want to see or use 7 The activation and installation fee to get service is too much 5 It’s not available where I live 2 I can access the Internet all I need to at work 1 None of above reasons 9 Combination of reasons 5 Other reason 5 Don’t know/refused 3 Source: Federal Communications Commission survey, October-November 2009; n=1,450 for nonusers. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 27
  • 28. OBI WOrkIng PaPer SerIeS nO. 1 As with nonusers, it is common for dial-up users to cite more fee, unwillingness to enter into a long-term contract—are next, than one reason that keeps them from using the Internet. Fifty- cited as most important by 28 percent of dial-up users. six percent cite at least three reasons and 18 percent cite five or Dial-up users who cited cost as a barrier then received a more (See Exhibit 20). follow-up question asking how much they would be willing to About half of dial-up users are satisfied with their service or pay for a broadband connection at home (See Exhibit 21). As are not heavy Internet users. Half also cite the monthly cost, with nonusers, responses fell into three categories. Among with a significant share saying the installation fee is a barrier. dial-up users citing monthly cost as a reason they do not have Nearly one-third say broadband is not available where they live. broadband: However, it is important to put this figure in proper context. ➤ 62 percent provided an estimate. The average amount First, dial-up users make up 6 percent of the adult population; this group was willing to pay for broadband was $28 per if 30 percent truly cannot get broadband where they live, this month and the median figure was $25. means just 2 percent of adults (from this set of respondents ➤ 25 percent said they did not know or refused to answer who received the question) lack broadband availability. Second, the question. since this is self-reported data on infrastructure availability, it ➤ 13 percent said they would not be willing to pay anything is possible that respondents are wrong as to whether broadband for broadband. is available where they live. Nonetheless, most (74 percent) of dial-up users who say availability is a barrier report that they Not-at-home users—6 percent of the adult population would get broadband if it were available in their area. Although the vast majority of Internet users have the means to When probed further on the most important barrier, issues go online from home, 6 percent do not. They use the Internet pertaining to the relevance of the service (that is, not using the from elsewhere—perhaps work, perhaps the library—but not Internet very much, being content with current dial-up service where they live. and not needing additional speed) together loom largest (38 For “not-at-home” users, worries about comfort with a percent) for dial-up users. Cost issues—monthly fee, activation computer or thinking there is nothing worthwhile online are not big barriers. Rather, affordability (e.g., monthly fee, activation Exhibit 20: The monthly cost is too expensive 50 Reasons Dial-Up Users, I’m happy with my current service 49 Who Make Up 6% of I do not use the Internet that much 44 the Adult Population, I do not want to enter into a long-term service contract 44 Have Not Switched to The activation or installation fee to get service is too much 44 Broadband (by Percentage) I do not need the additional speed it would offer 31 It’s not available where I live 29 Source: Federal Communications Commission survey, October-November 2009; n=567 for home dial-up users or those who did not know connection type. Exhibit 21: The monthly cost is too expensive 19 Reasons Dial-Up Users I do not use the Internet that much 18 Have Not Switched It’s not available where I live 17 to Broadband I’m happy with my current service 16 (by Percentage) I do not want to enter into a long-term service contract 4 I do not need the additional speed it would offer 4 The activation or installation fee to get service is too much 5 None of these reasons 7 Other/Don’t know 2 Source: Federal Communications Commission survey, October-November 2009; n=567 for home dial-up users or those who did not know connection type. 28 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 29. OBI WOrkIng PaPer SerIeS nO. 1 charge or computer cost) plays a larger role. While those reasons What non-adopters are Willing to pay are front and center, “not-at-home” users cite multiple reasons Across the three categories, 51 percent of all non-adopters why they are not online. Two-thirds (66 percent) list at least said monthly cost was a barrier, although fewer said it was the three reasons and 11 percent cite five or more (See Exhibit 22). primary reason they lack service. As is evident from the pre- Affordability clearly comes to the forefront when “not-at- ceding discussion, not all of those who said cost was a barrier home” users are asked about the most important barrier they were able to respond to follow-up questions on how much they face. Some 41 percent cite the monthly cost or the outlay for would be willing to pay for service. Across all three categories, a computer as the reasons they do not use the Internet from just more than half (52 percent) of those who said monthly cost home. was a barrier could provide an estimate to the nearest dollar of As with prior groups, “not-at-home” users who said monthly what they would pay for service. As a share of all non-adopters, cost was a barrier received a follow-up question on what they 26 percent provided an estimate of willingness to pay for would be willing to pay for Internet access at home. Among service. The willingness-to-pay figure for all non-adopters was “not-at-home” users for whom cost is a barrier (See Exhibit 23): $25 per month. ➤ 74 percent gave a dollar estimate; the average figure cited The $25 average does not mean that all non-adopters asked as this group’s willingness to pay for broadband was $21 the question would purchase service if they faced a price at that per month and the median figure was $20. level. Some respondents cited a figure above $25, many cited ➤ 16 percent said they did not know or refused to answer a figure below. In fact, among those who offered an answer, 65 the question. percent cited a figure of $20 per month or more for WTP. The ➤ 10 percent said they would not be willing to pay anything vast majority (91 percent) of those who gave an answer cited a for broadband. figure of $10 per month or more. Taking these respondents at their word—that they really would get broadband at the price they state—and representing Exhibit 22: Monthly cost is too expensive 57 Reasons “Not-at- The activation and installation fee to get service is too much 51 Home” Users Have Not I can access the Internet all I need to at work 34 Switched to Broadband I cannot afford a computer 32 (by Percentage) I am worried about all the bad things that can happen if I use the Internet 31 It’s not available where I live 15 There is nothing on the Internet I want to see or use 14 The Internet is just a waste of time 14 I am not comfortable using a computer 13 Source: Federal Communications Commission survey, October-November 2009; n=392 for not-at-home users. Exhibit 23: Monthly cost is too expensive 27 Reasons “Not-at- I cannot afford a computer 15 Home” Users Have Not I can access the Internet all I need to at work 11 Switched to Broadband I am worried about all the bad things that can happen if I use the Internet 9 (by Percentage) The activation and installation fee to get service is too much 6 It’s not available where I live 5 I am not comfortable using a computer 5 The Internet is just a waste of time 4 There is nothing on the Internet I want to see or use 2 None of these reasons 7 Source: Federal Communications Commission survey, October-November 2009; n=392 for not-at-home users. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 29
  • 30. OBI WOrkIng PaPer SerIeS nO. 1 the figures above for $20 per month and $10 per month as a exist for many current broadband users, some non-adopters are share of all adults: concerned enough to avoid the Internet altogether. This may ➤ A $20 offer would yield an increase in overall broadband have to do with the age of those citing digital literacy as a bar- adoption of 6 percentage points. rier. Their median age is 62. ➤ A $10 offer would yield an increase of 8 percentage points. relevance: 19 percent of non-adopters say they do not think digital content delivered using broadband is sufficiently com- These small impacts reflect the fact that the survey tested pelling to justify getting it. Many view broadband as an avenue the hypothesis that there are multiple reasons for non-adop- to irrelevant content, and others seem content with the offline tion, with cost being one among them. The survey found that alternatives currently available to them. These respondents say, people had multiple barriers with, as noted previously, 51 for instance, that the Internet is a “waste of time,” they do not percent citing cost as one of them, followed up by a question believe there is anything worth viewing online or, for dial-up on willingness-to-pay. Just more than half cited a figure, while users, they cite contentment with their current service. Each 28 percent answered “don’t know” and the balance said they of those reasons indicates that these non-adopters are on the would pay nothing for broadband. other side of a perceptual chasm with respect to broadband. Finally, it is important to underscore the uncertainties Unlike broadband users, they are not attuned to online con- inherent in measures of willingness-to-pay. Respondents are tent’s potential to provide information or opportunities for asked to make judgments on the value of a service that, many at learning. least, have used sparingly (if at all); they therefore lack infor- mation on service attributes, which is key to placing a dollar other: 11 percent of non-adopters cited a variety of reasons value. For that reason, measures of willingness to pay, and the that did not group into an identifiable category or offered no scenario on how much additional adoption would ensue from response. specific price offerings, should be taken as illustrative, not predictive. combination: 4 percent of non-adopters cited as their main reason a combination of reasons listed, whether that was price, summary on most important reason worries about “bad” things online or other items. for non-adoption The three baskets of non-adopters were asked similar kinds lack of availability: 5 percent of non-adopters said that of questions about why they do not have broadband or any the main reason they do not have the Internet or broadband is Internet service at home. Pulling the reasons together across that it is not available where they live. This response is highest all three groups of non-adopters yields the following consoli- among dial-up users, 21 percent of whom say that broadband is dated reasons for non-adoption among all those who do not not available where they live. have broadband at home. use high-speed at work: Finally, 3 percent of non-adopt- cost: 36 percent of non-adopters cite a cost-related reason ers said the main reason they do not have broadband at home as their main barrier to adoption, with 15 percent pointing to is because they use broadband as much as they want at the the monthly cost of service, another 10 percent citing afford- workplace. ability of a computer and 9 percent saying they do not want a For the “lack of availability” reason, it is also important to long-term service contract or find the installation fee too high. distinguish respondents citing lack of availability as the most The remaining 2 percent cited a combination of these reasons. important reason for not having broadband from those who cited it as one of several reasons. Because people could list digital literacy: 22 percent of non-adopters give a digital multiple reasons for not adopting (prior to being asked the literacy-related topic as their main barrier. These are non- most important reason), some who cited lack of infrastructure adopters who said that “they are not comfortable using a availability could (and did) not cite that as their most impor- computer” or they “are worried about all the bad things that tant reason for non-adopting. The upshot is that 12 percent of can happen on the Internet.” Digital literacy breaks down as non-adopters cited lack of availability as a reason for not hav- follows: 10 percent of non-adopters cite “worries about bad ing broadband, while 5 percent of non-adopters cited it as the things” and 12 percent said they are “not comfortable with com- most important reason. The 12 percent of non-adopters citing puters.” Although worries about online hazards (e.g., exposure lack of available infrastructure translates into 4 percent of all to inappropriate content or the possibility of identity theft) Americans who say they cannot get broadband because it is not 30 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 31. OBI WOrkIng PaPer SerIeS nO. 1 available where they live. Finally, it is important to underscore Among the 47 percent of non-adopters who listed digital that this is a self-reported assessment about the availability of literacy as a barrier: broadband infrastructure, not an actual measure of whether ➤ 70 percent also listed cost as a barrier. broadband is in the respondent’s neighborhood. ➤ 57 percent also listed lack of relevance as a barrier. When respondents can cite multiple reasons An important consequence of this analysis is that non- for not adopting adopters typically face multiple barriers to adoption. Even It is also worth summarizing, for all non-adopters, what share though cost leads the way, most of the time even those who cited any of the reasons listed as possible barriers. The survey worry about cost need help in other areas. Likewise, those who permitted non-adopters to list several reasons for not adopting say lack of relevance is a barrier also have issues with digital before asking them to state the most important reason. Because literacy and, to a less extent, cost. respondents could list more than one reason—in fact most cited multiple reasons—the figures below sum to more than segmenting the population of non-adopters 100 percent. Each of the three baskets of non-adopters reveals tendencies as to why people do not have broadband. Financial matters are cost: 66 percent of non-adopters cited any of the reasons critical for not-at-home Internet users, while dial-up users that constitute the “cost” category, that is, two-thirds of non- seem distant from the Internet, apparently not seeing a lot adopters listed monthly access fee, cost of computer, installa- online that might lure them to faster service. Nonusers—the tion fee, or reluctance in entering into a long term contract, as a largest group of non-adopters, constituting two-thirds of reason they do not use the internet or broadband. non-adopters—are harder to pin down. Many cite reasons sug- gesting they do not see the relevance of the Internet, yet some relevance: 52 percent of non-adopters said that they found point to costs and others worry about risks in being online. the internet to be a waste of time, they did not think there is Even if the survey responses yielded crisp reasons for non- anything worth seeing online, or they were content with their adoption across categories of nonbroadband users, it would be current service. worthwhile to know more than just a person’s stated barrier to adopting. In particular, it would help to know something about digital literacy: 47 percent of non-adopters listed their the person’s ability to clear the barrier. lack of comfort with computers or worries “about all the bad One technique to do this is to segment the population of things that can happen on the internet” as reasons for not hav- non-adopters into several groups based on measures of a re- ing broadband. spondent’s relationship to the Internet. The following analysis does this and it is built on two dimensions: lack of availability: 12 percent of non-adopters said that ➤ Proximity to Icts: All survey respondents were asked their inability to get service where they live is why they do not whether they have computers, cell phone or premium TV have broadband. at home. Nonbroadband adopters also were asked if they To underscore how barriers to access interact in users’ had ever used a computer. minds, the following shows the frequency with which those who ➤ attitudes about the Internet and computers: cited a particular barrier also pointed to others. Respondents received questions about their overall com- Among the 66 percent of non-adopters who cited cost as a fort with computers and attitudes about the Internet. barrier: ➤ 50 percent also listed a digital literacy-related barrier. The survey details the specific attitudinal questions and ➤ 48 percent also listed concerns about the relevance of the responses for broadband adopters and non-adopters. It is internet as a barrier. clear that people without broadband at home use ICTs. Among non-adopters: Among the 52 percent of non-adopters who listed lack of ➤ 80 percent have premium TV, either cable or satellite relevance as a barrier: television. ➤ 45 percent also listed a digital literacy-related barrier. ➤ 70 percent have a cell phone. ➤ 28 percent also listed on the cost-related reasons. ➤ On average, these cell phone users spend $73 per month on service. ➤ 49 percent say they use computers, at least occasionally. ➤ 42 percent have at least one working computer at home. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 31
  • 32. OBI WOrkIng PaPer SerIeS nO. 1 ➤ Among nonbroadband-adopting computer users, Group 2: Digital Hopefuls: 8 percent 34 percent say they are very comfortable using them. The Digital Hopefuls like the idea of being online, but lack the resources to connect using broadband. Members of this On the second issue, non-adopters have a mix of attitudes group have positive attitudes about how the Internet might about the Internet—some positive, others less so. Exhibit 24 be a tool for learning and personal productivity; 86 percent compares “strongly agree” responses for broadband adopters strongly believe it is a useful tool for learning and 68 percent and non-adopters. believe people can be more productive using the Internet. But The attitudinal and proximity questions serve as input into 22 percent cannot afford a computer (and only 3 percent have statistical analysis that classifies respondents based on their one at home), 15 percent cannot afford the monthly access fee responses. That analysis includes all non-adopters—dial-up, and only 9 percent have past experience using broadband. A not-at-home users and nonusers. The analysis yielded four significant share (28 percent) cites digital literacy as a barrier, groups of non-adopters, each with distinct characteristics the highest of any group. Given help—likely a lot of it—in clear- described below. The analysis also identifies the key adoption ing cost hurdles to access, members of this group have a positive barriers in each group, although answers to those questions probability of adopting broadband. were not inputs into the statistical analysis that created the Consistent with their lack of resources to connect, the groups.19 Digital Hopefuls are the least well off economically of any group. Some 44 percent live in households with annual incomes Group 1: Digitally Distant: 10 percent of the general of $20,000 or less. Nearly half (49 percent) have not gradu- population ated from high school. This group, with a median age is 57, has This group does not see the point of being online and does not the highest share of Hispanics (26 percent). One in five are have the computer skills to dive in even if its members were African-Americans and 50 percent are white. so inclined. Few (15 percent) in this group see the Internet as a tool for learning and most see it as a dangerous place, either Group 3: Digitally Uncomfortable: 7 percent for children (53 percent) or personal information (56 percent). The Digitally Uncomfortable has the equipment for access Very few (11 percent) have a computer. Many (25 percent) (nearly all have a computer), but the group has tepid attitudes cite digital literacy as a barrier to adoption, and nearly half about the usefulness of the Internet and low comfort levels are retired from the work force. Just 8 percent of this group with computers. Just 24 percent say they are very comfortable have any experience with broadband (i.e., they used at a place using them, and just more than one-third (37 percent) see the other than home or used to have home high-speed service). The internet as a productivity tool. Although some in the Digitally Digitally Distant group has a low probability of adoption in the Uncomfortable group could use a break on the monthly ac- near future. cess fee (16 percent), more need help learning about relevant The important demographic characteristics of this group content (18 percent) and improving their digital literacy (17 have to do with age. With a median age of 63, it is the oldest of percent). Twenty-one percent cite a lack of available infra- any of the groups; 47 percent are age 65 or older. The Digitally structure as a reason they do not have broadband. Distant also has the highest rate of retirees across the four As to attitudes, 53 percent worries about security of per- groups, at 45 percent. They are also not well off economically, sonal information online, although a majority (68 percent) with 50 percent reporting annual household incomes below sees the Internet as a useful tool for learning and information $30,000. gathering. The path to access and robust use might be long for this group. Because they have already cleared the computer ownership hurdle and because one-third (35 percent) have Exhibit 24: There is too much pornography and offensive material on the Internet 65 How Strongly Do You It is too easy for my personal information to be stolen online 57 Agree or Disagree The Internet is a valuable source for information and learning 59 With the Following It is important for children to learn how to use the Internet 54 Statements The Internet is too dangerous for children 46 (% of Non-adopters Who “Strongly Agree”) People can be more productive using the Internet 37 Source: Federal Communications Commission survey, October-November 2009; n=2,334 for non-adopters. 32 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 33. OBI WOrkIng PaPer SerIeS nO. 1 past experience with broadband, the Digitally Uncomfortable, affordability for the Digitally Adrift, monthly bill for the with the right support, stand a reasonable chance of adopting Digitally Disposed), but other factors are important, too. broadband. At the same time, although cost leads as a primary barrier From a socioeconomic perspective, members of the Digitally for non-adopters, nearly two-thirds cite a reason other than Uncomfortable group are very much the average non-adopter. cost. As noted earlier, non-adopters generally mention more The median age is 55 (versus 54 for all non-adopters), 63 per- than one obstacle for adoption—meaning interventions to lure cent are white (the same as the average), and 27 percent have people to broadband should be comprehensive in nature. Cost either graduated from college or at least attended college (also relief will work effectively for many non-adopters but only the average). when accompanied by training programs to bolster their digital skills and information about content that is relevant to their Group 4: Near Converts: 10 percent lives. Similarly, addressing digital literacy in isolation is not This group has many of the qualities of broadband adopters— likely to be effective unless cost is addressed and efforts are a high rate of computer ownership (76 percent use them oc- made to inform people about online content relevant to them. casionally and 68 percent own one), comfort with computers (56 percent describe themselves as very comfortable with the device) and positive attitudes about the Internet’s useful- v. fOCuS On key ness (three-quarters say it is a valuable learning tool). And 42 percent have used broadband in the past. However, nearly 41 POPulaTIOn grOuPS percent cite financial issues as the main barrier to home broad- band access, principally the monthly access fee. Concerns about the Internet’s relevance to their lives also play a role. At 10 per- Although this paper has referred to differences among popula- cent of the general population, this group has a high probability tion subgroups, this section has key figures for several groups of adopting broadband—if they clear financial constraints. of interest, as well as discussion of notable variations in behav- The Near Converts are much younger than other non-adopt- ior and attitudes for these groups. This focus on specific groups ers (their median age is 45) and they have much higher levels covers data on adoption and non-adoption. of educational attainment. Half have either graduated from college or attended college. Their incomes are in the middle families (respondent is the parent of child ranges (30 percent have annual household incomes between under 18 Who lives in household) $40,000 and $75,000) and most are employed full or part time. With the integration of information technology into so many Exhibit 25 summarizes the main barriers that each seg- dimensions of family life—whether is communicating with one ment of non-adopters faces. A clear implication of Exhibit 25 another or schools—it is no surprise that those with children is this: The greater the conversion probability, the larger role under age 18 in the house generally have a lot of information the monthly Internet bill plays as a barrier. The Near Converts technology. Nine out of 10 (87 percent) have Internet access, have the highest probability of becoming broadband adopt- and 74 percent have broadband at home. The survey inter- ers; the monthly access cost is the main hurdle. The most viewed parents, of whom 92 percent have a cell phone, and difficult-to-convert group is the Digitally Distant; they have many (39 percent) use it sometimes to access the Internet. no single reason standing out as a barrier. In between the two Among kids under 18 living at home, 45 percent have cell extremes, cost plays a role in slightly different ways (computer phones. For children between the ages of 12 and 17, some 75 percent have cell phones. Exhibit 25: Segment % as share of all Americans today Adoption barriers faced Summary of Most Cost 1. Near Converts 10 Important Barriers by (mostly monthly bill) Segment 2. Digital Hopefuls 7 Cost (mostly monthly bill), relevance Cost (mostly PC), 3. Digitally Uncomfortable 8 digital literacy, accessibility Cost, digital literacy, 4. Digitally Distant 10 relevance, accessibility Total Non-adopters 35 Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 33
  • 34. OBI WOrkIng PaPer SerIeS nO. 1 Although families with minor children at home have higher loW-income americans broadband adoption rates, there is still a gap between broadband There is a clear relationship between broadband adoption and adoption and computer ownership. Among such families, 87 per- income. Some 40 percent of those with household incomes of cent have a working computer at home—a 13-percentage-point $20,000 per year or less have broadband at home, while 91 per- gap compared with 74 percent of those without children who cent of those with household incomes above $75,000 per year have a working computer at home. have it (See Exhibit 27). The survey asked parents or guardians the ages of their Not only does broadband adoption vary by income, indi- children. When it came broadband adoption, there was little viduals with broadband at home exhibit some differences in difference between parents with children from ages 5 to 12, and behavior and attitude toward broadband depending on their those with children between 13 and 17; the level was 72 percent income. In general terms, lower-income broadband users for families with younger children and 75 percent for families are more likely to use their home high-speed connections to with older ones. To the extent large adoption differences unfold address important life issues, such as job searches or educa- by age, younger parents are less likely to have broadband at tion, and for entertainment. Higher-income broadband users, home than older ones. For parents with offspring at home in the however, are more likely than low-income ones to shop online, 18-to-29 age range, 65 percent have broadband at home, com- contact government and bank (See Exhibit 27). pared with 77 percent of parents older than that. When asked whether particular online activities were very As to online activities, parents tend to engage in activities important to them, low-income and upper-income broadband that focus on convenience and community. Fully 86 percent have users share the idea that communicating with family and used their broadband connection to keep up with community friends is very important. By a modest margin (74 percent to 68 news, 85 percent for online purchases and 57 percent to get percent), low-income broadband users are more likely to say advice or information from a government agency about a health that than upper-income ones. Low-income users are also more or safety issue; all of these exceed the average. When asked what likely to say that entertainment activities are very important to activities are very important to them, 44 percent cite keeping up them. One in five (21 percent) low-income broadband users say with news about the community; that is above the average. that watching TV, movies or videos online is very important to them, compared with 8 percent of upper-income users who say Non-adopters this. However, upper-income broadband users are nearly twice Cost—and the monthly fee in particular—comes front and as likely to say that shopping online is very important to them; center as the barrier to adoption for non-adopting families. 32 percent say this versus 17 percent of low-income broadband Half (48 percent) say that cost is the primary reason they are users. not online, with 24 percent specifically citing the monthly. Issues such as digital literacy (16 percent) or lack of relevance Non-adopters (13 percent) figure much less. Note that the “other” category Unsurprisingly, those with the lowest incomes are most likely in Exhibit 26, and ones that follow, refers to all other reasons to cite cost as the main barrier to having broadband at home cited by respondents, which may include items such as lack of (See Exhibit 28). availability and ability to use the Internet from work. For the other barriers—digital literacy and relevance—the stories are mixed. For digital literacy, aside from low-income Exhibit 26: Main reason cited for not having the Families with minor children at home Other non-adopters Barriers to Adoption: Internet or broadband Families with Minor Cost 46 31 Children at Home Digital literacy 16 24 (by Percentage) Relevance 13 21 Other 25 25 Source: Federal Communications Commission survey, October-November 2009. 34 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 35. OBI WOrkIng PaPer SerIeS nO. 1 people citing it less often, around one-fourth of non-adopters To an extent greater than average, African-Americans with- point to it as a primary barrier. The Internet’s perceived lack of out broadband at home use the mobile Internet as a substitute, relevance is more of an issue in the middle-income ranges. For but it is not a very widespread practice. Among African- higher-income non-adopters (of which there are few), hard-to- Americans who have gone online with their mobile device, categorize “other” reasons dominate. 78 percent have broadband at home; for all mobile Internet users, the home broadband adoption rate is 89 percent. For african-americans African-Americans without broadband at home, 20 percent For African-Americans, online access is, relative to the average, have used the Internet on their handheld devices; that figure is somewhat less oriented to home high-speed wireline access just 9 percent for all non-adopters. and more focused on mobile Internet. Nearly three in five (59 percent) African-Americans have broadband at home, but Another distinctive pattern among African-Americans is 39 percent have used the Internet on their mobile handheld that broadband adoption gaps are particularly acute for older device. For African-Americans, home broadband adoption African-Americans. In fact, broadband adoption for African- trails the national average by six percentage points; for mobile Americans is 76 percent in the 18–29 age group, essentially the Internet use, African-Americans outpace the national average same others. Thereafter, however, African-Americans are much by nine percentage points. less likely than average to adopt broadband at home when the Exhibit 27: Low-income broadband users Upper-income broadband users By Income, Reasons (annual household income (annual household income of $20K or less) of $75K or more) for Broadband Use (by Percentage) Activities in which low-income users lead Get information about a job or apply for one 77 60 Play games 70 44 Stream music to your computer 64 55 Take a class for credit 31 25 Play complicated role-playing games 21 9 Activities in which upper-income users lead Buy a product 74 93 Go to local, state, or government Web site 68 86 Get international news 64 85 Banking 51 80 Review a product or service 49 62 Contact a government agency 42 62 about health or safety issue Source: Federal Communications Commission survey, October-November 2009. Exhibit 28: Cost Digital literacy Relevance Other Reasons for Not Under $20K 47 17 15 21 Adopting Broadband $20K–$30K 37 23 18 23 (by Percentage) $30K–$40K 38 21 14 27 $40K–$50K 29 24 23 25 $50K–$75K 25 23 26 23 Over $75K 25 20 12 43 Source: Federal Communications Commission survey, October-November 2009. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 35
  • 36. OBI WOrkIng PaPer SerIeS nO. 1 focus is on age and the gap is most pronounced among senior home; this compares to an average 36 percent among non- citizens (See Exhibit 29). adopters (See Exhibit 32). Because of the high share of young people in the ranks of Dissecting the cost factor shows that 15 percent of African- African-American broadband adopters, the online activities Americans point to the affordability of a personal computer of African-American broadband users takes on a youthful (above the 10 percent average), and 15 percent cite the level of cast. Some 83 percent of African-Americans with broadband the monthly bill (right at the average). have used the connection to search for or apply for a job, far above the 57 percent rate for other broadband users. African- hispanics American broadband users are also more likely to take a class The gap separating Hispanics’ home broadband use from for credit online or participate in a blog (See Exhibit 30). the average is wider than the comparable figure for African- As to what they see as very important, African-American Americans; 49 percent of Hispanics have broadband at home. broadband users are very likely to say communicating with There is also a gap in adoption within the groups, as Spanish- family and friends, although at a somewhat lower rate than speaking Hispanics are far less likely to have broadband at average. African-Americans are, however, more likely to say home than those whose facility with English is such that they that the Internet is very important to them for keeping up with opted to take the survey in English. For Hispanic respondents community news and entertainment (See Exhibit 31). who chose to take the survey in Spanish, just 20 percent have broadband at home. Hispanics who took the survey in English Non-adopters are right at the 65 percent average for home high-speed adop- For African-Americans without broadband access at home, cost tion (See Exhibit 33). is cited as the main reason—and it weighs in at a rate somewhat The adoption gap for Hispanics is most acute among young higher than average. Forty-two percent of African-Americans adults. Some 57 percent of Hispanics between the ages of 18 and cite cost as the primary reason they do not have broadband at 29 have broadband, more than 20 points below the average. Just Exhibit 29: African-Americans All others in sample Broadband Adoption by All 59 66 Age: African-Americans (by Percentage) Ages 18–29 76 75 Ages 30–49 67 75 Ages 50–64 49 66 Ages 65+ 21 36 Source: Federal Communications Commission survey, October-November 2009. Exhibit 30: African-Americans All other broadband adopters Online Activities: Get information about or apply for a job 83 57 African-Americans Update a blog (either own or group blog) 37 24 (by Percentage) Take a class toward a degree 37 22 Source: Federal Communications Commission survey, October-November 2009. Exhibit 31: % who say this is very important to them African-Americans All other broadband adopters Online Attitudes: Communicating with family & friends 61 67 African-Americans Keeping up with news in the community 51 38 Watching TV shows, movies or other video 18 9 Source: Federal Communications Commission survey, October-November 2009. 36 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 37. OBI WOrkIng PaPer SerIeS nO. 1 more than half (53 percent) of those between the ages of 30 and ➤ Streaming or downloading music: Consistent with the 49 have broadband, 25 points behind the average. youthful Hispanic population, they are more likely to Hispanics are not laggards when it comes to the cell phone. download or stream music—69 percent versus the 49 per- The vast majority (85 percent) has one and, along with African- cent average. Americans, members of this group are active users of the At the same time, the broadband connection is also im- mobile Internet; 39 percent have taken advantage of online portant for Hispanics for following what is going on in the access with their handheld. To a somewhat greater extent than community. Fully 80 percent use the Internet for getting local African-Americans, mobile access takes the place of broadband or community news, and 52 percent identify this kind of activ- at home. One-quarter of Hispanics who do not have broadband ity as very important to them, compared with the 38 percent at home access the Internet using their mobile device. Among figure for others. Hispanics who have used the Internet via handheld devices, 68 percent have broadband at home, another indication that the Non-adopters mobile Internet fills the wireline void for some Hispanics. The story among Hispanic non-adopters is cost. Half cite As to online activities, Hispanics stand out, relative to the it as the main barrier to adoption. Twenty-one percent average, in two clear ways. (See Exhibit 34) say the monthly online bill is too much for ➤ Job Search: Hispanics are more likely to use their them, and 23 percent say they cannot afford a computer. high-speed connection to search for or apply for a job—68 percent versus the 59 percent average. Exhibit 32: Main reason cited for not having African-Americans Other non-adopters Barriers to Adoption: the Internet or broadband African-Americans Cost 42 35 (by Percentage) Digital literacy 20 22 Relevance 13 20 Other 25 24 Source: Federal Communications Commission survey, October-November 2009. Exhibit 33: Hispanics All others in sample Broadband Adoption All 49 67 by Age: Hispanics (by Percentage) Ages 18–29 57 81 Ages 30–49 53 78 Ages 50–64 32 67 Ages 65+ 23 36 Source: Federal Communications Commission survey, October-November 2009. Exhibit 34: Main reason cited for not having Hispanics Other non-adopters Barriers to Adoption: the Internet or broadband Hispanics Cost 52 32 (by Percentage) Digital literacy 18 22 Relevance 14 20 Other 16 26 Source: Federal Communications Commission survey, October-November 2009. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 37
  • 38. OBI WOrkIng PaPer SerIeS nO. 1 people With disabilities On the whole, however, broadband-using people with dis- As noted earlier, 23 percent of respondents in this survey abilities do fewer things online than average. The survey asked answered “yes” to one of six questions intended to identify an half of all respondents about eight different online activities, individual as having a disability. In the survey, those with a dis- and the other half about another eight activities. Among both ability were older than most respondents, with a median age of sets of activities, respondents with disabilities and broadband 57. This explains, in part, the lower overall Internet penetration were less active than the average. For one set, broadband-using rate for this group (56 percent) and the low rate of home high- people with disabilities did 4.6 activities on average compared speed adoption (42 percent). Although 74 percent have a cell to 5.1 for broadband users without a disability. For the other phone, people with disabilities are not very likely to access the set, the figures were 3.3 and 3.7, respectively, for broadband us- Internet with it (only 18 percent have). ers with disabilities and other broadband users. The differences One-third of respondents with a disability were over the may not seem great, but they indicate a level of online activity age of 65 in the survey, and only 25 percent have broadband at for broadband users with disabilities that is about 12 percent home. At the other end of the spectrum, one-third is under the lower than that of broadband users who are not disabled. age of 50, and 57 percent has broadband at home—more than People with disabilities share have many of the character- 20 points below average. It is notable that senior citizens with istics that are often associated with lower levels of Internet disabilities are 76% less likely to have broadband than seniors use—advanced age and lower income. However, having a dis- who do not have a disability. (See Exhibit 35) ability—even when those other characteristics are taken into Notable among those with disabilities is the narrower scope account—is significantly correlated with lower levels of online of online activities. For some activities, people with disabili- activity. ties who have broadband are nearly as likely to engage in them as the overall average. Three quarters (73 percent) of people Non-adopters with disabilities use their broadband connection to get news People with disabilities do not differ greatly from the average about their communities, not too far from what other broad- in terms of why they do not have broadband at home. They are band users say (81 percent). People with disabilities are also somewhat more likely to say they cannot afford a computer (15 within range of others for use of their broadband connection to percent cite it as the main reason versus the 10 percent aver- visit a government Web site (71 percent versus the 80 percent age), which may be due to the extra expense for some users average). for fitting a computer with the capability to allow them to use broadband (See Exhibit 36). Exhibit 35: People with disabilities All others in sample Broadband Adoption All 42 72 by Age for People with a Disability Ages 18–29 59 78 (by Percentage) Ages 30–49 56 77 Ages 50–64 43 72 Ages 65+ 25 44 Source: Federal Communications Commission survey, October-November 2009. Exhibit 36: Main reason cited for not having People with Disabilities Other non-adopters Barriers to Adoption: the Internet or broadband People with Disabilities Cost 37 35 (by Percentage) Digital literacy 25 19 Relevance 17 19 Other 21 27 Source: Federal Communications Commission survey, October-November 2009. 38 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 39. OBI WOrkIng PaPer SerIeS nO. 1 rural americans residents between 18 and 29 have broadband at home, com- In the survey, rural Americans are people living in counties that pared with the average of 75 percent. do not contain any portion of a metropolitan statistical area It appears that some rural high-speed users take advantage (MSA). This is distinct from urban areas, defined as central of the distance-bridging capacity of broadband. More than cities in MSAs, and suburban areas, which are any portion of three-quarters (77 percent) of rural broadband users have an MSA county that is not a central city. Place of residence bought a product online. Although this is below the average is typically easy to determine from the respondents’ landline 84 percent rate for other broadband users, it is notable in light telephone number. Cell phone numbers may not map as eas- of the fact that rural Americans have a higher share of two ily to place of residence, and for that reason, respondents are qualities that discourage online buying: age and lower incomes. asked to name the zip code for where they live. This is used to Similarly, 25 percent of rural Americans have taken a class on- determine place of residence (rural, urban, or suburban) for line—something more the province of young people—matching respondents contacted on their cell phone. Using these two the 24 percent rate for other broadband users. methods, 17 percent of respondents met the definition of resi- As to what is important to them, 60 percent of rural broad- dents of rural America. band users say communicating with family and friends is very There is a sizable gap between rural Americans with any sort important, followed by 33 percent who value keeping up with of access to the Internet (71 percent) and those with broad- community news and 33 percent who say sharing content with band at home (50 percent). Ten percent of rural residents have others is very important. dial-up Internet connection, above the average, reflecting the difficulty in running broadband infrastructure to some corners Non-adopters of rural America (See Exhibit 37). Fully 80 percent of rural The reasons rural Americans cite for not having broadband at Americans have cell phones, but just 20 percent have used home differ modestly from the average—with one exception. them for Internet access. Rural non-adopters are twice as likely as urban or suburban Rural Americans tend to be older than their suburban non-adopters to say broadband is not available where they live and urban counterpart (the median age for rural adults is 50 (by a 10 percent to 4 percent margin, which is reflected in the compared with 46 for all American adults) and with somewhat “other” category below). They are less likely than other non- lower incomes. Younger rural Americans are far behind when it adopters to say cost is a barrier (31 percent to 38 percent) and comes to broadband adoption by age; 56 percent of rural relatively few (9 percent) say that the expense of a computer keeps them from broadband (See Exhibit 38). Exhibit 37: Rural residents All others in sample Broadband Adoption All 50 68 by Geographic Location (by Percentage) Ages 18–29 56 78 Ages 30–49 63 76 Ages 50–64 51 67 Ages 65+ 29 37 Source: Federal Communications Commission survey, October-November 2009. Exhibit 38: Main reason cited for not having Rural residents National Average Barriers to Adoption: the Internet or broadband Rural Americans Cost 31 38 (by Percentage) Digital literacy 23 21 Relevance 19 18 Service not available 10 4 Other 18 19 Source: Federal Communications Commission survey, October-November 2009. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 39
  • 40. OBI WOrkIng PaPer SerIeS nO. 1 Appendix: Demographic and socio-economic overview of those with various technological assets Tables on (as a percentage of Americans) Demographics Cable or Landline Cell Working com- and Use satellite TV phone phone puter at home Gender Male 48 46 50 50 Female 52 54 50 50 Parents with minor children at home 33 33 34 36 Those who report they have a disability 23 25 20 18 18–29 20 15 24 24 30–49 34 34 37 38 50–64 27 29 26 26 65+ 17 20 13 12 White (not Hispanic) 71 74 71 73 Black (not Hispanic) 11 11 12 10 Hispanic (English or Spanish speaking) 12 10 12 11 Less than high school 13 12 10 7 High school graduate 34 35 34 32 Some college 25 24 26 28 College + 29 29 30 34 Under $20K 15 14 15 12 $20K–$30K 9 9 9 8 $30K–$40K 9 8 10 10 $40K–$50K 9 9 9 10 $50K–$75K 14 14 15 16 $75K–$100K 11 11 12 13 Over $100K 15 15 15 17 Don’t know/refused 18 20 17 16 Urban 29 28 30 29 Suburban 50 52 49 51 Rural 17 18 16 17 Number of cases 4,301 4,373 4,268 3,650 Source: Federal Communications Commission survey, October-November 2009. 40 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 41. OBI WOrkIng PaPer SerIeS nO. 1 Appendix: Demographic & socio-economic overview of those who do various activities online (% broadband adapters in group) Visit local, state or Use a Submit a Get local or Upload federal Download Buy a social review for Play community or share government or stream product network- a product games news* content* Web site* music* online* ing site* or service* online* Male 78 44 79 54 82 52 53 41 Female 82 52 78 49 84 58 57 56 Parents with minor children at 86 51 82 58 85 59 56 53 home Those who report they have a 73 41 71 45 73 47 57 52 disability 18–29 86 67 75 75 86 85 57 63 30–49 85 49 80 58 84 57 58 50 50–64 74 37 83 34 84 38 53 38 65+ 58 26 68 17 69 16 39 27 White 79 48 81 49 87 54 55 46 (not Hispanic) Black 80 48 70 52 69 56 63 63 (not Hispanic) Hispanic (English or 80 45 74 70 70 59 49 49 Spanish speaking) Less than high 75 48 60 57 64 55 51 76 school High school 73 36 66 48 74 50 47 47 graduate Some college 81 48 83 53 84 58 56 54 College+ 85 56 87 53 92 57 60 40 Under $20K 80 46 68 64 74 58 49 70 $20–30K 78 50 78 56 64 57 55 58 $30–40K 78 43 77 58 76 48 51 54 $40–50K 80 51 70 48 84 53 57 38 $50–75K 75 42 80 42 87 54 52 45 $75–$200K 84 49 87 52 92 60 54 45 Over $100K 86 60 85 58 95 61 69 43 Don’t know/ 73 40 72 42 73 45 48 43 refused Number of cases 1,378 1,378 1,378 1,378 1,378 1,378 1,378 1,378 Source: Federal Communications Commission survey, October–November 2009. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 41
  • 42. OBI WOrkIng PaPer SerIeS nO. 1 Appendix: Demographic & socio-economic overview of those who do various activities online (% broadband adapters in group) Get infor- Get advice Play com- Get inter- mation from a govern- plicated national or about or Post to own ment agency Download Do any Take a role playing national apply for blog or about a health or stream banking class games news* a job group blog or safety issue video online online online Male 70 60 23 52 46 69 21 18 Female 78 65 27 57 44 73 28 14 Parents with minor children at 78 65 27 57 44 73 28 14 home Those who report they have a 66 50 24 49 41 58 19 21 disability 18–29 80 83 45 52 66 72 38 23 30–49 80 65 26 55 43 75 27 16 50–64 75 47 16 56 29 64 13 7 65+ 64 14 6 44 18 55 6 7 White 77 55 24 53 40 71 22 11 (not Hispanic) Black 72 83 37 55 54 57 37 31 (not Hispanic) Hispanic (English or 78 68 27 56 47 68 25 22 Spanish speaking) Less than high 65 64 22 43 38 46 19 17 school High school 67 60 25 42 43 59 17 18 graduate Some college 77 61 27 57 42 70 30 18 College+ 87 58 26 62 43 78 25 9 Under $20K 64 77 30 43 47 51 32 22 $20–30K 70 63 29 45 36 62 28 21 $30–40K 76 64 33 59 48 64 22 19 $40–50K 79 53 23 45 36 68 13 13 $50–75K 76 56 25 52 48 77 27 15 $75–$200K 86 66 26 55 52 77 25 14 Over $100K 84 56 24 67 49 82 26 5 Don’t know/ 73 50 20 50 36 57 18 15 refused Number of cases 1,293 1,293 1,293 1,293 1,293 1,293 1,293 1,293 Source: Federal Communications Commission survey, October–November 2009. 42 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 43. OBI WOrkIng PaPer SerIeS nO. 1 Appendix: Demographic and socioeconomic overview of non-adopters by selected barriers (% of those facing barrier, by demographic) Cost Digital Literacy Relevance Lack of availability Male 40 45 43 49 Female 60 55 57 51 Parents with minor children at home 32 17 15 33 Those who report they have a disability 41 46 37 21 18–29 24 6 10 18 30–49 29 22 21 30 50–64 26 28 22 35 65+ 19 44 44 16 Median Age 47 61 61 50 White (not Hispanic) 54 65 71 78 Black (not Hispanic) 16 13 9 11 Hispanic (English or Spanish speaking) 27 16 14 5 Less than high school 34 29 27 10 High school graduate 42 51 47 38 Some college 14 12 15 33 College+ 9 8 11 20 Under $20K 38 24 24 22 $20–30K 15 15 14 15 $30–40K 10 10 7 10 $40–50K 7 9 10 7 $50–75K 5 8 11 5 $75–$200K 3 4 2 3 Over $100K 2 2 3 2 Don’t know/refused 21 28 30 21 Urban 37 28 24 8 Suburban 38 44 47 42 Rural 21 26 25 46 Number of cases 738 516 445 148 Source: Federal Communications Commission survey, October–November 2009. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 43
  • 44. OBI WOrkIng PaPer SerIeS nO. 1 Appendix: Demographic & socio-economic overview of non-adopter segments (% of non-adopters) Digitally Digitally Distant Digital Hopefuls Uncomfortable Near Converts Male 37 42 43 50 Female 63 58 57 50 Parents with minor children at home 18 22 25 28 Those who report they have a disability 49 51 35 24 18–29 10 13 17 21 30–49 18 22 22 37 50–64 25 26 27 30 65+ 47 38 34 12 Median Age 63  57 55  45  White (not Hispanic) 66 50 63 71 Black (not Hispanic) 11 20 10 13 Hispanic (English or Spanish speaking) 18 26 22 10 Less than high school 36 49 30 5 High school graduate 48 38 43 45 Some college 9 9 17 28 College+ 6 4 10 22 Under $20K 36 44 27 13 $20–30K 14 15 22 9 $30–40K 8 8 12 12 $40–50K 5 3 8 14 $50–75K 3 3 7 16 $75–$200K 1 1 3 8 Over $100K 1 0 3 8 Don’t know/refused 32 26 20 22 Urban 25 20 27 25 Suburban 43 35 44 43 Rural 29 39 26 27 Number of cases 649 450 478 757 Source: Federal Communications Commission survey, October–November 2009. 44 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 45. OBI WOrkIng PaPer SerIeS nO. 1 Appendix: Non-adopter segments’ attitudes towards the Internet and technological assets (% of non-adopters) Digitally Digitally Distant Digital Hopefuls Uncomfortable Near Converts The Internet is dangerous for children 53 48 43 40 It is important for child to learn 16 76 63 66 to use Internet The Internet is valuable for info/learning 15 86 68 74 There is too much offensive or 62 68 65 65 questionable material on the Internet People can be more productive with 0 68 37 49 Internet It is too easy for personal 56 66 53 54 information to be stolen online Number of cases 649 450 478 757 Source: Federal Communications Commission survey, October–November 2009. Appendix: Information and Communication Technologies by Segment (% of segment) Digitally Digitally Distant Digital Hopefuls Uncomfortable Near Converts Cable TV 59 56 46 61 Satellite TV 27 26 37 33 Cable OR Satellite TV 78 75 75 86 Desktop computer 9 2 83 55 Laptop computer 3 1 38 32 Desktop OR laptop computer 11 3 97 68 Cell phone 52 56 76 93 Number of cases 649 450 478 757 Source: Federal Communications Commission survey, October–November 2009. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 45
  • 46. OBI WOrkIng PaPer SerIeS nO. 1 were made to contact every sampled cellular telephone num- meThOdOlOgy ber. Sample was released for interviewing in replicates, which are representative subsamples. Using replicates to control the Broadband consumer Survey release of sample helps ensure that complete call procedures Prepared by Princeton Survey Research Associates are followed. Calls were staggered over times of day and days of Internationalfor the Federal Communications Commission the week to maximize the chance of making contact with poten- December 2009 tial respondents. Interviewing was spread as evenly as possible across the field period. summary For the landline sample, interviewers asked to speak with The Broadband Service Capability Survey, sponsored by the the adult in the household who had the most recent birthday. Federal Communications Commission, obtained telephone If the selected adult was not at home or could not complete the interviews with a nationally representative sample of 5,005 interview at the time of contact, interviewers arranged to call adults living in the United States. The survey was conducted by back the selected person at a later time. For the cellular sample, Princeton Survey Research International. Interviews were con- interviews were conducted with the person who answered the ducted in English and Spanish by Princeton Data Source LLC phone. Interviewers verified that the person was an adult and from October 19 to November 23, 2009. The data was weighted in a safe place before administering the survey. Cellular sample to correct known demographic discrepancies. The margin of respondents were offered a post-paid cash reimbursement for sampling error for the complete set of weighted data is ±1.6 their participation. percentage points. The interviewing was done in two phases that ran concur- Details on the design, execution and analysis of the survey rently. The first phase included interviews with all adults while are discussed below. the second phase obtained interviews with an oversample of adults who were not broadband adopters (i.e., the non-adopter design and data collection procedures oversample). Respondents who were contacted in phase two and were adopters were screened out. Sample Design Two samples were used for data collection—a random digit Weighting and analysis dial (RDD) landline sample and an RDD cell sample. The Weighting is generally used in survey analysis to compensate landline sample frame was an equal probability sample across for sample designs and patterns of non-response that might all active blocks.20 All blocks within a county were sorted in bias results. The data was weighted to correct for three sample ascending order by area code, exchange and block number. A elements that could potentially bias survey estimates—[1] the sampling interval was computed for each county in our sample oversampling through screening of additional non-adopters, by summing all eligible blocks in the county and dividing that [2] different probabilities of selection based on the number sum by the quota assigned to the county. From a random start of adults in each household and the number of telephones between zero and the sampling interval, blocks were systemati- that each respondent has access to and [3] disproportionate cally selected from each county. Once a block was selected, a non-response. two-digit random number was appended to the block to create a phone number. Business numbers were not excluded at the Non-adopters Oversample sampling stage. Rather they were flagged during sampling and A sample adjustment was made to account for the non-adopters purged before dialing. Additionally, protected numbers were oversample. This adjustment simply adjusted the proportion not excluded from the sample frame.21 of non-adopters in our total sample to match the proportion The cellular sample was not list-assisted because no list of of non-adopters in the first phase of interviewing. This adjust- cellular numbers exists. Rather, cellular phone numbers were ment was made individually for the landline and cell samples. systematically sampled from dedicated wireless 100-blocks Table 1 shows the oversample adjustment (OSADJUST) that and shared service 100-blocks with no directory-listed landline was made to the data. numbers. Different Probabilities of Selection Contact Procedures We made two sample adjustments to the data that address Interviews were conducted from October 19 to November unequal selection probabilities. One adjustment accounts for 23, 2009. As many as 15 attempts were made to contact every within household clustering and is a function of the number of sampled landline telephone number and as many as 7 attempts adults in each household while the second adjustment accounts 46 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 47. OBI WOrkIng PaPer SerIeS nO. 1 for the overlapping landline and cell sample frames and the number of landline telephones in each household. (p1 + p2+ p3) for respondents with three or more The Probability of Selection Adjustment (PSA) corrects for 3p1 + 2p2+ p3 phones the fact that respondents in the landline sample have different probabilities of being sampled depending on how many adults live in the household. Since we only sample one person per Differential Non-response household, adults who live with no other adults have a greater The final step in weighting the data consisted of raking sample chance of being sample than adults who live with one or more demographics to match population parameters. The sample other adults. was balanced—by form—to match national population param- To compute the PSA, first define n1 as the number of people eters for sex, age, education, race, Hispanic origin, region (U.S. in the landline sample who live in single-adults households and Census definitions), population density, and telephone usage. n2 as the number of people in the landline sample that live in The basic weighting parameters came from a special analysis multi-adult households. The PSA equals: of the Census Bureau’s 2009 Annual Social and Economic Supplement (ASEC) that included all households in the United n1 + n2 for landline respondents in single-adult States. The population density parameter came from an analy- households sis of Census 2000 data. The telephone usage parameter came n1 + 2n2 from an analysis of the most recently available National Health Interview Survey data.22 2(n1 + n2) for landline respondents in multiple- Raking was accomplished using Sample Balancing, a spe- n1 + 2n2 adult households cial iterative sample weighting program that simultaneously balances the distributions of all variables using a statistical technique called the Deming Algorithm. An input weight equal 1 for cellphone respondents to the product of OSADJUST, PSA and PUA was used for the raking. Weights were trimmed to prevent individual interviews The Phone Use Adjustment (PUA) corrects for different from having too much influence on the final results. The use probabilities of selection based on the number of landline of these weights in statistical analysis ensures that the demo- phones in each respondent’s household and whether or not the graphic characteristics of the sample closely approximate the respondent has a cell phone. demographic characteristics of the national adult population. To compute the PUA, first define p1 as the number of respon- Table 2 compares weighted and unweighted sample demo- dents with only one phone, p2 as the number of respondents graphic distributions to population parameters. with two phones and p3 as the number of respondents with three or more phones, the PUA equals: Effects of Sample Design on Statistical Inference Post-data collection statistical adjustments require analysis 3(p1 + p2+ p3) procedures that reflect departures from simple random sam- for respondents with one phone pling. PSRAI calculates the effects of these design features so 3p1 + 2p2+ p3 that an appropriate adjustment can be incorporated into tests of statistical significance when using these data. The so-called 2(p1 + p2+ p3) “design effect” or deff represents the loss in statistical effi- for respondents with two phones 3p1 + 2p2+ p3 ciency that results from disproportionate sample designs and Table 1: Main sample Main sample + oversample Adjustment (OSADJUST) Oversample Adjustment Landline Non-adopters 36.2% 48.6% 0.745 Adopters 63.8% 51.4% 1.241 Cellular Non-adopters 21.2% 37.7% 0.562 Adopters 78.8% 62.3% 1.265 Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 47
  • 48. OBI WOrkIng PaPer SerIeS nO. 1 systematic non-response. The total sample design effect for n this survey is 1.37. n ∑ w i2 PSRAI calculates the composite design effect for a sample of i=1 size n, with each case having a weight, wi as: ( ) deff = formula 1 n 2 ∑ wi i=1 Table 2: Parameter Unweighted Sample Weighted Sample Sample Demographics Gender Male 48.5 45.1 48.1 Female 51.5 54.9 51.9 Age 18–24 12.6 7.4 12.1 25–34 17.8 10.6 16.8 35–44 18.2 13.9 17.8 45–54 19.6 20.2 19.5 55–64 15.1 19.2 15.0 65+ 16.6 26.3 16.8 Education Less than high school 14.1 10.0 13.0 High school graduate 34.7 33.3 34.1 Some college 24.1 22.8 24.3 College + 27.1 33.1 28.0 Race/Ethnicity White (not Hispanic) 68.8 74.1 68.8 Black (not Hispanic) 11.5 9.2 11.1 Hispanic 13.7 8.7 12.6 Other (not Hispanic) 6.0 6.6 6.1 Region Northeast 18.5 17.6 18.3 Midwest 22.0 25.8 22.5 South 36.8 37.5 37.0 West 22.7 19.1 22.1 County Pop. Density 1—Lowest 20.1 24.7 20.5 2 20.0 22.5 20.4 3 20.1 20.5 20.4 4 20.2 17.7 20.0 5—Highest 19.6 14.7 18.6 Phone Use LLO 13.6 14.7 13.6 Dual users 65.6 72.8 66.4 CPO 20.8 12.6 20.1 48 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 49. OBI WOrkIng PaPer SerIeS nO. 1 In a wide range of situations, the adjusted standard error in 95 out every 100 samples drawn using the same methodol- of a statistic should be calculated by multiplying the usual ogy, estimated proportions based on the entire sample will be formula by the square root of the design effect (√deff ). Thus, no more than 1.6 percentage points away from their true values the formula for computing the 95% confidence interval around in the population. The margin of error for estimates based a percentage is: on non-adopters is ±2.4 percentage points. It is important to ( ) remember that sampling fluctuations are only one possible ˆ ˆ p (1- p ) ˆ p± √deff x 1.96 √ n formula 2 source of error in a survey estimate. Other sources, such as respondent selection bias, questionnaire wording and reporting inaccuracy, may contribute additional error of greater or lesser ˆ where p is the sample estimate and n is the unweighted magnitude. number of sample cases in the group being considered. The survey’s margin of error is the largest 95% confidence response rate interval for any estimated proportion based on the total Table 3 reports the disposition of all sampled telephone sample—the one around 50%. For example, the margin of error numbers ever dialed from the original telephone number for the entire sample is ±1.6 percentage points. This means that samples. The response rate estimates the fraction of all eligible Table 3: Landline—combined Cell—combined Sample Disposition Total Numbers Dialed 75974 33914 Non-residential 6069 618 Computer/Fax 3602 34 Cell phone 61 Other not working 39194 12847 Additional projected not working 3330 868 Working numbers 23718 19547 Working Rate 31.2% 57.6% No Answer / Busy 1110 289 Voice Mail 2989 4730 Other Non-Contact 123 12 Contacted numbers 19496 14516 Contact Rate 82.2% 74.3% Callback 2372 2278 Refusal 11661 8310 Cooperating numbers 5463 3928 Cooperation Rate 28.0% 27.1% Language Barrier 213 78 Child's cell phone — 1105 Adopter screen-out 1580 1169 Eligible numbers 3670 1576 Eligibility Rate 67.2% 40.1% Break-off 188 53 Completes 3482 1523 Completion Rate 94.9% 96.6% Response Rate 21.9% 19.4% Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 49
  • 50. OBI WOrkIng PaPer SerIeS nO. 1 respondents in the sample that were ultimately interviewed. At PSRAI it is calculated by taking the product of three compo- aCknOWledgemenTS nent rates:23 ➤ Contact rate—the proportion of working numbers where The author wishes to thank several members of the National a request for interview was made24 Broadband Taskforce for their help in making this report a ➤ Cooperation rate—the proportion of contacted numbers reality. Foremost is Ellen Satterwhite who provided invaluable where a consent for interview was at least initially ob- help on a number of fronts—from insights on survey questions, tained, versus those refused to analysis of the data, to editorial judgment. Jessica Strott also ➤ Completion rate—the proportion of initially cooperating participated in design of the survey instrument, and her ideas and eligible interviews that were completed greatly contributed to its quality. The survey was carried out by Princeton Survey Research Thus the response rate for the landline sample was 22 per- Associates International and the entire team there—Evans cent. The response rate for the cellular sample was 19 percent. Witt, Margie Engle, and Jonathan Best—were instrumental in designing and carrying out this survey. They have the author’s sincere thanks. 50 F e d e r a l c O m m u n I c at I O n S c O m m I S S I O n | W W W. B r O a d B a n d.g O V
  • 51. OBI WOrkIng PaPer SerIeS nO. 1 endnOTeS 1 Because it is difficult to tell whether such access meets 15 As in the example above, the findings that the correlation the standards to qualify as broadband, users of wireless between skills and number of activities is the result of an handheld access are not counted as home broadband OLS regression in which the dependent variable is the users—though most wireless handheld internet users number of activities a user has done and the independent qualify as home broadband users. variables include gender, race, education, income and 2 Mobile wireless broadband users are counted as home the number of tech terms the user reports understanding broadband users if that is their source of home access. “very well.” 3 1 percent of non-adopters could not be classified into any 16 Cody Barbierri, “Smartphone usage continues to grow of the three groups of non-adopters. in US, according to Forrester Research.” Demo Beat, 4 National Telecommunications and Information Admin- http://demo.venturebeat.com/2010/01/05/smartphone- istration, Digital Nation: 21st Century America’s Progress usage-continues-to-grows-in-us-according-to-forrester- Toward Universal Broadband Internet Access. February research/. Accessed on February 12, 2010. 2010. Available online at: http://www.ntia.doc.gov/ 17 John B. Horrigan, Wireless Internet Use, Pew Internet & reports/2010/NTIA_internet_use_report_Feb2010.pdf. American Life Project, July 22, 2009. 5 More information is available at the FCC’s Web site: 18 According to the Census, 54 million people have a http://www.fcc.gov/form477/. disability, including speech, hearing, vision, mobility 6 Lee Rainie et.al., The Ever Shifting Internet Population. or cognitive disorders. U.S. Census Bureau, Americans Pew Internet & American Life Project, April 2003. With Disabilities: 2005, Current Population Reports, 3, Available online at: http://www.pewInternet.org/ 4 (2008), http://www.census.gov/Press-Release/www/ Reports/2003/The-EverShifting-Internet-Population- releases/archives/income_wealth/013041.html. A-new-look-at-Internet-access-and-the-digital-divide. 19 In constructing the segments, a statistical cluster aspx, accessed on January 18, 2010. analysis was performed only on those respondents 7 John B. Horrigan. If you build it, will they log on? Pew who did not have broadband at home. Several different Internet & American Life Project, January 2009. Avail- cluster solutions were evaluated for their effectiveness able online at: http://www.pewInternet.org/~/media// in producing cohesive groups that were distinct from one Files/Reports/2009/PIP_Broadbandpercent20Barriers. another and substantively meaningful. The final solution pdf, accessed on January 18, 2010. selected was judged to be strongest on a statistical basis 8 Ofcom, Accessing the Internet at Home, June 2009. and to be most persuasive from a substantive perspec- Available online at: http://www.ofcom.org.uk/research/ tive. telecoms/reports/bbresearch/bbathome.pdf, accessed 20 Active blocks are defined as 100 contiguous phone on January 18, 2010. numbers (e.g., 609-924-9200 to 609-924-9299) with at 9 John B. Horrigan, How Americans Get in Touch with least one residential directory listings. Government, Pew Internet & American Life Project, May 21 Phone numbers in SSI random-digit database are 2004. Available online at: http://www.pewInternet.org/ flagged as “protected” if they have recently been pulled Reports/2004/How-Americans-Get-in-Touch-With- for any sample order. While excluding protected num- Government.aspx, accessed on January 18, 2010. bers results in a “fresher” sample (i.e., one that consists 10 Omnibus Broadband Initiative (OBI) internal analysis of of numbers that have not been recently pulled), we feel various industry sources. that excluding these numbers would potentially bias our 11 Respondents were asked simply to estimate their sample. monthly bill to the nearest, but not prompted to make 22 Blumberg SJ, Luke JV. Wireless substitution: Early that estimate exclusive or inclusive of taxes or fees that release of estimates from the National Health Interview may be imposed on service. Survey, July-December, 2008. National Center for 12 Internal OBI analysis of data from TNS Telecoms and Health Statistics. May 2009. Telogical Systems. 23 PSRAI’s disposition codes and reporting are consistent 13 The finding that the correlations are statistically sig- with the American Association for Public Opinion nificant is derived from an ordinary least squares (OLS) Research standards. regression in which the dependent variable is the num- 24 PSRAI assumes that 75 percent of cases that result in a ber of activities a user has done and the independent constant disposition of “No answer” or “Busy” are actu- variables include gender, race, education, income, level ally not working numbers. of online skills (as discussed in the next section) and the specific attitudinal variable of interest. 14 Eszter Hargittai, “An Update on Survey Measures of Web-Oriented Digital Literacy.” Social Science Computer Review, Volume 27, no. 1, February 2009, pp. 130–137. Federal cOmmunIcatIOnS cOmmISSIOn | Br OadB and adOP tIOn and uSe In amerIca 51